The aim of this Innovative Training Network is to train a new generation of creative, entrepreneurial and innovative early stage researchers (ESRs) in the research area of measurement and estimation of signals using knowledge or data about the underlying structure.

With its combination of ideas from machine learning and sensing, we refer to this research topic as “Machine Sensing”. We will train all ESRs in research skills needed to obtain an internationally-recognized PhD; to experience applying their research a non-Academic sector; and to gain transferable skills such as entrepreneurship and communication skills.

We will further encourage an open “reproducible research” approach to research, through open publication of research papers, data and software, and foster an entrepreneurial and innovation-oriented attitude through exposure to SME and spin-out Partners in the network. In the research we undertake, we will go beyond the current, and hugely popular, sparse representation and compressed sensing approaches, to develop new signal models and sensing paradigms. These will include those based on new structures, non-linear models, and physical models, while at the same time finding computationally efficient methods to perform this processing.

We will develop new robust and efficient Machine Sensing theory and algorithms, together methods for a wide range of signals, including: advanced brain imaging; inverse imaging problems; audio and music signals; and non-traditional signals such as signals on graphs. We will apply these methods to real-world problems, through work with non-Academic partners, and disseminate the results of this research to a wide range of academic and non-academic audiences, including through publications, data, software and public engagement events.

MacSeNet is funded under the H2020-MSCA-ITN-2014 call and is part of the Marie Sklodowska-Curie Actions — Innovative Training Networks (ITN) funding scheme

Check out the videos our ESRs have made about their research:

Workshops

SPARS2017 5th-8th June 2017

In collaboration with the SpaRTaN project, the network bid and successfully won the right to host SPARS 2017 at IST in Lisbon, Portugal. This is a large workshop, 150-200 attendees, which is well established within a relevant field. On the first day of this event there was a showcase of the two MSCA ITN projects. This was achieved in two ways, first there was an allocated slot in the oral programme for the Network Co-ordinator, Prof. Mark Plumbley, to explain the ethos of the MCSA programme and the 2 projects that have been funded. Secondly, there was a special poster session for the Fellows to present the research work they had been doing during the project.

You can find out more about the workshop on the SPARS2017 webpage.

Sparse Representations and Compressed Sensing Workshop 23rd March 2018

This one-day workshop, organised in collaboration with the SpaRTaN Initial Training Network, will include invited keynote talks by Karin Schnass (Universität Innsbruck, Austria) and Jean-Luc Starck (CEA-Saclay, France), oral presentations and posters. The talks and posters will include theoretical advances in sparse representations, dictionary learning and compressed sensing, as well as advances in areas such as brain imaging and MRI, hyperspectral imaging, audio and visual signal processing, inverse imaging problems, and graph-structured signals.

You can find out more about the workshop on the webpage.

Scientific Training Events

Spring School 2016 held in Fraunhofer IDMT, Ilmenau, Germany 4th-8th April 2016.

The SpaRTaN-MacSeNet Spring School on Sparse Representations and Compressed Sensing Spring School was aimed at graduate students, researchers and industry professionals working in this fast moving and exciting area. The five day school was split into two components, during three days, a panel of experts offered lectures and tutorials covering the theory of sparse representations, compressed sensing and related topics; alongside applications of these methods in areas such as image processing, audio signal processing, and signal processing on graphs. The remaining two days were devoted to software carpentry, giving researchers the computing skills they need to get more done in less time and with less pain

Find out more details at the Spring School Webpage.

Second Summer School 2017 held in Lisbon, Portugal, 31st May-2nd June 2017

We held our Second Summer School for scientific training in Lisbon, Portugal. We co-located the event with the SPARS Workshop as this enabled us to secure great speakers and get a wide audience from outside the network. The courses consisted of the following speakers and topics:

In addition to these speakers there was a poster session for the Fellows and external participants to practise presenting their posters prior to SPARS and a panel discussion where several of our tutors answered questions from the tutees about the current field and where they think it will head in the future. The panel was chaired by Prof. Mário Figueiredo and questions were taken both on the day and in advance via a post-it-note board.

Find out more details on the Summer School Webpage.

Research Skills Training Events

The network has held 2 training weeks covering transferable skills for researchers and all our ESRs have attended one of these weeks. The first was in Surrey in September 2015 and the second preceded the Ilmenau Spring School in April 2016. Both training weeks focussed on transferable skills aligned to the Vitae Researcher Development Framework and were provided by SURREY’s Researcher Development Team who as well as providing courses for SURREY’s researchers have had experience running similar events for other ITNs and SEPNet (the South East Physics Network).

We wanted to cover topics which were important to all researchers as well as some of the more specific areas that are likely to come up for our highly mobile ESRs. The theme of the first sessions of both training weeks reflected this by covering unwritten rules, those between different cultures caused by different countries and languages as well as those caused by different stages of career or different education experiences. The other sessions covered the skills researchers need to plan and develop their career, they were formally introduced to the Vitae Researcher Development Framework and were encouraged to look at the skills a researcher needs. In week one this day included an employer panel discussion which allowed the ESRs to talk to senior researchers or managers in SMEs, International Companies, Start-ups and Academia. The panellists were asked to come with an idea of what they looked for when recruiting researchers so that the ESRs could ask questions about what it was like to work in various environments and what skills were specific or common across employers. Following this the ESRs were able to use the Action Planning tools to help them identify areas where they needed further training. During the second training week the employer panel was swapped for networking skills to enable them to make the most of the industry and academic visitors at the summer school that was about to follow. Other sessions looked at the skills researchers need to present their research and engage with different audiences and Dr Sophie Wehrens, the MacSeNet Ethics Advisor, prepared a session on understanding ethics in research with examples for the ESRs to discuss and analyse to give them an insight into how ethics applies to their research.

Second Training Week

From the 16th - 23rd November 2016 we ran the Second Training Week at EPFL in Lausanne, Switzerland. The first three days focussed on Science 2.0 topics including science communication, open access, publishing and data management. It was followed up by a three-day sandpit where the Fellows worked on small collaborative projects and put into practice the tools and skills they had learnt during the previous training weeks.

Third Training Week

The first week in April 2017 saw our Third Training Week, hosted in Edinburgh, UK. Timed to coincide with the Edinburgh International Science Festival the aim of the week was communication and entrepreneurship. Sessions covered innovation and entrepreneurship, graphic design and data visualisation, communicating to the press, intellectual property, licencing, financing and pitching. The week was ended by a morning with a voice coach and a public engagement session on the streets of Edinburgh.

Project Co-ordinator

Prof. Mark Plumbley, Centre for Vision, Speech and Signal Processing, University of Surrey, UK.

Project Administrator

Dr Helen Cooper, Centre for Vision, Speech and Signal Processing, University of Surrey, UK.

Contact

helen.cooper@surrey.ac.uk

Recruiting Is Now Closed

MacSeNet is an EU-funded, Marie Sklodowska-Curie Innovative Training Network, bringing together leading academic and industry groups to train a new generation of interdisciplinary researchers in efficient Machine Sensing theory and algorithms.

There are 18 Marie Sklodowska-Curie Early Stage Researcher (ESR) positions, which allow the researcher to work towards a PhD. The ESRs will be recruited to start mid 2015 for a duration of 36 months.

Each ESR will be working on an independent personal project and will have secondments linked to their research to other partners in the network, the planned secondments are listed below but may change as the individual projects evolve. They will also attend ITN progress meetings and Training events throughout Europe and possibly conferences and events internationally.

Marie Sklodowska-Curie ESRs are paid a competitive salary which is adjusted for their host country. Please see the individual positions below to find the annual salary for that host country (figures are given in Euros prior to employer and employee tax being deducted). Since the ESR positions include secondments to other hosts and for the researcher to move countries the EU also provides a Mobility Allowance, this is higher for researchers who have a family (family is defined as persons linked to the researcher by (i) marriage, or (ii) a relationship with equivalent status to a marriage recognised by the national legislation of the country of the beneficiary or of the nationality of the researcher, or (iii) dependent children who are actually being maintained by the researcher).

ESRs should be within four years of the diploma granting them access to doctorate studies at the time of recruitment.

In addition, to be eligible for a position as a Marie Sklodowska-Curie Early Stage Researcher you must not have spent more than 12 months in the host country in the 3 years prior to starting.

In order to be sure that you meet these requirements please fill in the eligibility form and include it with your application.

(Position Filled) ESR1 : Robust unsupervised learning (FRANCE) Apply by 2015-05-01

Location: INRIA/CNRS/ENS Paris, France

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €41,425

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: June to Sept 2015

Duration: 36 months

Closing Date for Applications: 2015-05-01

The aim of this PhD project is to develop new robust unsupervised learning methods suitable for the needs identified in other disciplines and by industrial partners. Several approaches to unsupervised learning exist; we plan to follow our earlier work on matrix factorization approaches, with a particular focus on robustness. This is both an optimization and a statistical problem, as the usual matrix factorization approaches are (a) based on non-convex optimization and (b) have several hyperparameters which may impact performance. The two problems are often treated separately. For example, Bayesian approaches will provide elegant solutions to the adaptivity of hyperparameters but put little emphasis on provable computational and stability issues, while research on convex relaxations often ignores the statistical issues related to hyperparameters and the automatic adaptivity of model to data. Combining these two approaches should bring the best of both worlds.

The 36 month project will be undertaken within the Computer Science Department of Ecole Normale Superieure, located in downtown Paris, within the INRIA/CNRS/ENS project-team SIERRA.

Planned secondments

  • Scientific: University of Edinburgh, UK, 3 months, for training on new structural models.
  • Cross-Sector: VisioSafe, Switzerland, 3 months for training on application of the new methods on image & video tracking.

Requirements

The applicant must have a Masters degree in Computer Science or Applied Mathematics (or any equivalent diploma), with knowledge of convex optimization, machine learning, and numerical linear algebra. 

Applications

To apply for the position, please provide:

  1. a letter of motivation including a maximum 1-page statement of your research interests, relevant skills and experience;
  2. a CV including publication list;
  3. your completed eligibility form.
  4. names and contact details of three referees willing to write confidential letters of recommendation.

All materials should be emailed as a single PDF file (<5 Mb) to Francis Bach with ‘PhD application MacSeNet ESR1′ in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR2 : Non-linear adaptive sensing/learning (FRANCE) Apply by 2015-05-01

Location: INRIA/CNRS/ENS Paris, France

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €41,425

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: June - Sept 2015

Duration:

Closing Date for Applications: 2015-05-01

In recent years, non-linear features and measurements have shown great empirical promises in machine learning and signal processing. The aim of this PhD project is to develop new methodologies adapted to non-linear measurements and predictors, with a particular focus on (a) methods based on convex reformulation of neural network training, (b) computationally efficient methods adapted to large-scale problems typically found in applications, and (c) principled adaptivity to the learning capacity of the underlying problems.

The 36 month project will be undertaken within the Computer Science Department of Ecole Normale Superieure, located in downtown Paris, within the INRIA/CNRS/ENS project-team SIERRA.

Planned secondments

  • Scientific: EPFL, Switzerland, 3 months, for training in structured signal models.
  • Cross-Sector: Songquito, Germany, 3 months to explore applications to musical audio analysis for games.

Requirements

The applicant must have a Masters degree in Computer Science or Applied Mathematics (or any equivalent diploma), with knowledge of convex optimization, machine learning, and numerical linear algebra.

Applications

To apply for the position, please provide:

  1. a letter of motivation including a maximum 1-page statement of your research interests, relevant skills and experience;
  2. a CV including publication list;
  3. your completed eligibility form.
  4. names and contact details of three referees willing to write confidential letters of recommendation.

All materials should be emailed as a single PDF file (<5 Mb) to Francis Bach with ‘PhD application MacSeNet ESR2′ in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR3 : Beyond sparse representations: efficient structured representations (UK)

Location: Institute for Digital Communications, School of Engineering, University of Edinburgh, UK

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €44,896

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: £25,275 pa. Mobility allowance: £5,249 pa. Family Allowance: £4,374 pa

Start Date: From July 2015

Duration: 36 Months

Closing Date for Applications: 2015-03-31

One PhD position under a European Union Marie Curie Initial Training program (H2020-MSCA-ITN-2014) MacSeNet is available in the Institute for Digital Communications at the University of Edinburgh, UK. The selected candidate will develop new efficient signal representations for use in signal processing, compressed sensing and machine learning. The candidate will specifically develop new mathematical models and algorithms that go beyond current sparse representation signal models.

The current use of sparse representations have proved very effective at improving the performance of many signal processing and machine learning tasks and led to the development of the new field of compressed sensing. However, such signal models have their limitations. This project will develop a new generation of signal models that retain the excellent efficiency and dimensionality reduction properties of sparse representations while also adding the capability to model invariances, nonlinearities and/or physical laws.

The Early Stage Researcher on this project will benefit from the partnership between the University of Edinburgh, INRIA Paris and the other academic institutions and industrial partners within the MacSeNet project. They will attend initial training events and be exposed to the research activities of all participants at regular six monthly progress meetings. They will engage in training events and secondments to other project partners.

Informal enquires are welcome and should be made to Prof. Mike Davies.

Planned secondments

  • Scientific: INRIA, France, 3 months, for training on structured norms.
  • Cross-Sector: VisioSafe, Switzerland, 3 months, for training in applications of methods in image and video analysis.

Requirements

Highly motivated, excellent candidates should ideally hold a valid Masters degree with a specialisation in machine learning and/or signal processing and should be eligible for immediate admission on the PhD programme at the University of Edinburgh.

Applications

Application is via the university job pages, vacancy reference : 032651

Please ensure your application includes your completed eligibility form and a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position;

Direct link to this advert (right click and copy link).

(Position Filled) ESR4 : Next generation compressed sensing techniques for quantitative MRI (UK)

Location: Institute for Digital Communications, School of Engineering, University of Edinburgh, UK

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €44,896

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: £25,275 pa. Mobility allowance: £5,249 pa. Family Allowance: £4,374 pa

Start Date: From July 2015

Duration: 36 Months

Closing Date for Applications: 2015-03-31

One PhD position under a European Union Marie Curie Initial Training program (H2020-MSCA-ITN-2014) MacSeNet is available jointly at the Institute for Digital Communications and the Brain Research Imaging Centre at the University of Edinburgh, UK. The selected candidate will study a new generation of compressed sensing (CS) techniques for accelerated acquisition in MRI. The candidate will specifically develop new mathematical models, pulse sequences and reconstruction algorithms for quantitative MRI.

Magnetic Resonance Imaging has already shown itself to be an ideal candidate for the application of compressed sensing theory. Excellent image reconstruction has been shown to be possible from undersampled k-space measurements through the application of compressed sensing principles and algorithms. However the real challenge and benefits lie in tackling advanced MR imaging techniques, such as quantitative MRI, MR spectroscopy and diffusion tensor imaging. These techniques require very long acquisition times and can suffer from bad motion artefacts induced during acquisition. These problems go beyond traditional CS solutions, and to tackle them will require the development of new structural signal models and sampling strategies.

The Early Stage Researcher will benefit from the partnership between the University of Edinburgh, GE Global Research, the Technical University of Munich and other academic institutions and industrial partners within the MacSeNet project. They will attend initial training events and be exposed to the research activities of all participants at regular six monthly progress meetings. They will engage in training events and secondments to other project partners.

Informal enquires are welcome and should be made to Prof. Mike Davies.

Planned secondments

  • Scientific: CTI, Greece, 3 months, for training in kernel and nonlinear methods.
  • Cross-Sector: GE Global Research, Germany 5 months, for training and experience on MRI scanning.

Requirements

Highly motivated, excellent candidates should ideally hold a valid Masters degree with a specialisation in medical imaging and/or signal processing and should be eligible for immediate admission on the PhD programme at the University of Edinburgh.

Applications

Application is via the university job pages, vacancy reference : 032650

Please ensure your application includes your completed eligibility form and a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position;

Direct link to this advert (right click and copy link).

(Position Filled) ESR5 : Next generation compressed sensing techniques for a fast and data-driven reconstruction of multi-contrast MRI (GERMANY) Apply By 2015-04-30

Location: Department of Informatics, Technical University Munich, Germany

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €36,872

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: From May 2015

Duration:36 Months

Closing Date for Applications: 2015-04-30

One PhD position is available jointly at the Institute of Informatics at the Technical University Munich (TUM) and GE Global Research (GEGR). The selected candidate will study the next generation of compressed sensing (CS) techniques for accelerated acquisition in MRI. In more detail, the candidate will develop accelerated pulse sequences for MR spectroscopy in the brain, derive mathematical models of the underlying physical principles and implement CS techniques for advanced reconstruction.

Magnetic Resonance Imaging has already shown itself to be an ideal candidate for the application of compressed sensing theory. Excellent image reconstruction has been shown to be possible from undersampled k-space measurements through the application of compressed sensing principles and algorithms. However the real challenge and benefits lie in tackling advanced MR imaging techniques, such as quantitative MRI, MR spectroscopy and diffusion tensor imaging. These techniques require very long acquisition times and can suffer from bad motion artefacts induced during acquisition. These problems go beyond traditional CS solutions, and to tackle them will require the development of new structural signal models and sampling strategies.

The Early Stage Researcher on this project will benefit from the partnership between TUM, GEGR, the University of Edinburgh and the other academic institutions and industrial partners within the MacSeNet project. They will attend initial training events and be exposed to the research activities of all participants at regular six monthly progress meetings. They will engage in training events and secondments to other project partners.

Informal enquires are welcome and should be made to Prof. Björn Menze.

Planned secondments

  • Scientific: University of Edinburgh, UK, 3 months, to explore ideas on compressed MRI.
  • Cross-Sector: GE Global Research, Germany, 1 day/week, for training and experience on MRI scanning.

Requirements

Highly motivated, excellent candidates should ideally hold a valid Master’s degree with a specialisation in biomedical engineering, physics, applied mathematics, medical imaging and/or signal processing and should be eligible for immediate admission on the PhD programme at the TUM.

Applications

Please ensure your application includes your completed eligibility form and a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position;

Direct link to this advert (right click and copy link).

(Position Filled) ESR6 : Next generation compressed sensing techniques for the fast and dynamic MRI (GERMANY)

Location: Department of Informatics, Technical University Munich, German

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €36,872

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: From March 2015

Duration: 36 Months

Closing Date for Applications: 2015-04-30

One PhD position is available jointly at the Institute of Informatics at the Technical University Munich (TUM) and GE Global Research (GEGR). The selected candidate will study the next generation of compressed sensing (CS) techniques for accelerated acquisition in MRI. More specifically, the candidate will develop work on accelerated pulse sequences for MR perfusion in the brain, develop fast reconstruction models of the 4D signal, and work towards a fast estimation of the local and global blood flow patterns.

Magnetic Resonance Imaging has already shown itself to be an ideal candidate for the application of compressed sensing theory. Excellent image reconstruction has been shown to be possible from undersampled k-space measurements through the application of compressed sensing principles and algorithms. However the real challenge and benefits lie in tackling advanced MR imaging techniques, such as MR spectroscopy, diffusion tensor imaging, or ASL perfusion imaging. These techniques require very long acquisition times to obtain sufficient a signal-to-noise ratio, as well as non-linear models for signal analysis. These problems go beyond traditional CS solutions, and to tackle them will require the development of new sampling strategies and structural signal models.

The Early Stage Researcher on this project will benefit from the partnership between TUM, GEGR, the University of Edinburgh and the other academic institutions and industrial partners within the MacSeNet project. They will attend initial training events and be exposed to the research activities of all participants at regular six monthly progress meetings. They will engage in training events and secondments to other project partners.

Informal enquires are welcome and should be made to Prof. Björn Menze

Planned secondments

  • Scientific: University of Edinburgh, UK, 3 months, to explore ideas on compressed MRI.
  • Cross-Sector: GE Global Research, Germany, 1 day/week, for training and experience on MRI scanning.

Requirements

Highly motivated, excellent candidates should ideally hold a valid Master’s degree with a specialisation in computer science, applied mathematics, a related engineering discipline, or physics and should be eligible for immediate admission on the PhD programme at the TUM.

Applications

Please ensure your application includes your completed eligibility form and a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position;

Direct link to this advert (right click and copy link).

(Position Filled) ESR7 : Blind source separation of functional dynamic MRI signals via distributed dictionary learning (GREECE) Apply by 2015-06-20

Location: University of Athens/Computer Technology Institute, Athens, Greece

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €34,590

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: €23,000 pa excluding Mobility and Family allowances

Start Date: April to September 2015

Duration: 36 Months

Closing Date for Applications: 2015-06-20

Brain tasks involving action, perception, cognition, etc., are performed via the simultaneous activation of a number of brain regions, which are engaged in proper interactions in order to effectively execute the task. In functional Magnetic Resonance Imaging (fMRI), brain activity is captured by detecting associated changes in blood flow within the brain. The obtained data stream comprises a mixture of the source signals which carry the valuable information required by the neuroscientists in understanding the brain functions.

Extracting information from fMRI data commonly relies on simplifying assumptions, which can compromise the accuracy of the analysis. In this PhD project, the candidate will develop and analyze new methods for fMRI data analysis and brain signal separation, based on non-linear data-driven approaches inspired from Machine Learning and Data Analysis, such as kernel-based dictionary learning. Emphasis will be put on low complexity and distributed solutions.

The candidate will benefit from the partnership with the other academic institutions and industrial partners within the MacSeNet project. She/he will attend training events and be exposed to the research activities of all participants at regular progress meetings. Moreover, she/he will engage in secondments to other project partners.

Informal enquires are welcome and should be sent to Prof Sergios Theodoridis.

Planned secondments

  • Scientific: TUM, Germany, 3 months, for training on fMRI processing.
  • Cross-Sector: Bioiatriki, Greece, 5 months, for training on fMRI data acquisition and experimental design.

Requirements

Candidates should hold a M.Sc. degree (or equivalent), in electrical engineering, computer science and/or engineering, applied mathematics, or related areas, with a strong background on linear algebra. Experience in machine learning and optimization will be greatly appreciated. Candidates are also expected to have good scientific programming skills, preferably in Matlab and/or C, a very good level of English, and availability to travel within the MacSeNet network.

Applications

To apply for the position, please provide:

  1. a letter of motivation (maximum one page) explaining how your research interests, skills and experience are relevant to the position.
  2. a detailed CV.
  3. your completed eligibility form.
  4. names and contact details of three referees willing to write confidential letters of recommendation.

All materials should be emailed as a single PDF file (<5 Mb) to Prof Sergios Theodoridis with ‘PhD application MacSeNet ESR7′ in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR8 : Functional neuroimaging data characterisation via tensor representations (GREECE) Apply by 2015-06-20

Location: University of Athens/Computer Technology Institute, Athens, Greece

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €34,590

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: €23,000 pa excluding Mobility and Family allowances

Start Date: April to September 2015

Duration: 36 Months

Closing Date for Applications: 2015-06-20

Brain tasks involving action, perception, cognition, etc., are performed via the simultaneous activation of a number of brain regions, which are engaged in proper interactions in order to effectively execute the task. In functional Magnetic Resonance Imaging (fMRI), brain activity is captured by detecting associated changes in blood flow within the brain. The obtained data stream comprises a mixture of the source signals which carry the valuable information required by the neuroscientists in understanding the brain functions.

Extracting information from fMRI data commonly relies on simplifying assumptions (e.g., linearity) and is based on matrix-based approaches that fail to exploit the inherent multi-way structure of brain data. In this PhD project, the candidate will investigate tensor (multi-way) models and associated algorithms that are adapted to the fMRI problem structure and assess their performance as compared to matrix-based schemes. Appropriate tensor factorization and tensor dictionary learning methods (both linear and nonlinear) will be developed and studied. Emphasis will be put on memory- and computation-efficient solutions.

The candidate will benefit from the partnership with the other academic institutions and industrial partners within the MacSeNet project. She/he will attend training events and be exposed to the research activities of all participants at regular progress meetings. Moreover, she/he will engage in secondments to other project partners.

Informal enquires are welcome and should be sent to Prof Sergios Theodoridis.

Planned secondments

  • Scientific: University of Edinbough, UK, 3 months, for training on multidimensional biomedical data.
  • Cross-Sector: Bioiatriki, Greece, 5 months, for training on fMRI data acquisition and experimental design.

Requirements

Candidates should hold a M.Sc. degree (or equivalent), in electrical engineering, computer science and/or engineering, applied mathematics, or related areas, with a strong background on linear algebra. Experience in machine learning and optimization will be greatly appreciated. Candidates are also expected to have good scientific programming skills, preferably in Matlab and/or C, a very good level of English, and availability to travel within the MacSeNet network.

Applications

To apply for the position, please provide:

  1. a letter of motivation (maximum one page) explaining how your research interests, skills and experience are relevant to the position.
  2. a detailed CV.
  3. your completed eligibility form.
  4. names and contact details of three referees willing to write confidential letters of recommendation.

All materials should be emailed as a single PDF file (<5 Mb) to Prof Sergios Theodoridis with ‘PhD application MacSeNet ESR8′ in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR9 : Phase imaging via sparse coding in the complex domain (PORTUGAL) Apply by 2015-06-15

Location: Instituto de Telecomunicações, Portugal

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €33,252

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: July to September 2015

Duration: 36 Months

Closing Date for Applications: 2015-4-15 or 2015-06-15 (if not filled)

Phase imaging is a class of inverse problems aimed at inferring phase images from nonlinear (often sinusoidal) and noisy measurements. It appears, for example, in interferometric synthetic aperture radar and sonar, 3D shape from structured light, magnetic resonance imaging, X-ray phase-contrast imaging, phase retrieval, phase-shifting interferometry, and sharing interferometry.

The measurements in the phase imaging systems are usually noisy and periodic functions of the original phase, as they are extracted from or periodic signals or waves. The periodic nature of the measurement process yields very difficult inverse problems, regardless the framework adopted to formulate and solve them. For example, under the regularization framework, and even using convex regularizes for phase, the presence of a periodic data term in the objective function leads to very hard non-convex optimization problems.

Objectives:

  1. Reformulate the phase imaging problem as a sparse regression in the complex domain. More specifically, the objective is the estimation of the complex exponential of the phase. Two techniques will be researched. The first is based on dictionary learning via matrix factorization with sparsity constraints on the code (i.e., the regression coefficients). The second is inspired on the BM3D (Block Matching 3 Dimensional filtering) method. The proposed research line differs, however, from BM3D in two aspects: firstly, it is applied to the complex-valued variables; secondly, it uses the higher order singular value decomposition (HOSVD) to design orthonormal transformations of spectral representations of the complex-valued phase.
  2. Develop phase unwrapping inference strategies to estimate the absolute phase from the complex exponentials (which are equivalent to phase modulo-2pi) obtained in objective 1). The phase unwrapping may be carried out over the estimates produced by the sparse regression methods developed in 1) or conceived to operate simultaneously with those regression methods.

Planned secondments

  • Scientific: Tampere University of Technology, Finland, 6 months, to explore phase unwrapping inference.
  • Cross-Sector: Noiseless Imaging, Finland, 3 months, for training on practical fast and effective implementation of methods.

Requirements

Candidates should hold an MSc degree in Electrical and/or Computer Engineering, Computer Science, or Applied Mathematics. The required skills include, with a flexible balance, advanced programming skills (C/C++, Matlab), strong background in statistics, signal processing and/or optimization, a good knowledge of English, and a strong motivation to work as part of a team.

Applications

To apply for the position, please provide:

  1. a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position
  2. a CV including publication list
  3. names and contact details of three referees willing to write confidential letters of recommendation
  4. your completed eligibility form

All materials should be emailed as a single PDF file to: bioucas@lx.it.pt with ‘PhD application MacSeNeT ESR9' in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR10 : Patch-based, non-local, and dictionary-based methods for blind image deblurring (PORTUGAL) Apply by 2015-06-15

Location: Instituto de Telecomunicações, Portugal

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €33,252

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: July to September 2015

Duration: 36 Months

Closing Date for Applications: 2015-4-15 or 2015-06-15 (if not filled)

Image deblurring (namely deconvolution) is one of the core and oldest problems in image processing. The state of the art methods for this task are based on non-local patch-based approaches, such as the well-known BM3D (Block Matching 3 Dimensional) filtering. BM3D, which achieves state-of-the-art results in image denoising and non-blind deconvolution, is a nonlocal image modelling technique based on adaptive, non-local, groupwise models. Of course, BM3D is not the only non-local image model with excellent performance in image restoration; another important class of methods that can be considered as belonging to the patch-based family includes dictionary based techniques. These techniques are based on modelling image patches as sparse combinations of elements of some dictionary, which is learned from training data or from the image to be restored.

Objectives:

The goal of this project is to extend the applicability of this type of methods (namely BM3D and dictionary-based methods) to the more realistic scenario of blind deconvolution. This is a challenging task, since the current versions of these methods rely on knowledge about the observation operator (blur kernel), but its success would have a high impact in the field of blind image restoration.

Planned secondments

  • Scientific: Tampere University of Technology, Finland, 6 months, to explore alternative blind deblurring approaches.
  • Cross-Sector: Noiseless Imaging, Finland, 3 months, for practical fast and effective implementation of blind deblurring.

Requirements

Candidates should hold an MSc degree in Electrical and/or Computer Engineering, Computer Science, or Applied Mathematics. The required skills include, with a flexible balance, advanced programming skills (C/C++, Matlab), strong background in statistics, signal processing and/or optimization, a good knowledge of English, and a strong motivation to work as part of a team.

Applications

To apply for the position, please provide:

  1. a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position
  2. a CV including publication list
  3. names and contact details of three referees willing to write confidential letters of recommendation
  4. your completed eligibility form

All materials should be emailed as a single PDF file to: mario.figueiredo@lx.it.pt with ‘PhD application MacSeNeT ESR10' in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR11 : Sparse coding in the complex domain for phase retrieval and lensless coherent diffractive imaging (FINLAND) Apply by 2015-06-20

Location: Tampere University of Technology, Finland

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €43,515

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: August to September 2015

Duration: 36 Months

Closing Date for Applications: 2015-5-1 or 2015-06-20 (if not filled)

One Early-Stage Researcher position is available. The selected candidate will study the advanced sparse approximation techniques for coherent diffractive imaging. Phase retrieval is a special noninterferometric technique for phase reconstruction from intensity measurements from single and multiple sensors. Lensless coherent diffractive imaging uses intensity measurements of multiple diffraction patterns collected with a localized illumination probe from overlapping regions. The fields of the potential applications include holography, microscopy, astronomy, bio- and medical imaging. In more detail, the coherent (phase) imaging is based on operation with complex-valued variables and requires development of novel methods and algorithms.

Planned secondments

  • Scientific: Instituto de Telecomunicações, Portugal, 5 months, for training on phase retrieval inverse problems.
  • Cross-Sector: Noisless Imaging, Finland, 3 days/month, for practical aspects of fast and effective implementation of the sparse coding strategies.

Requirements

Candidates should hold a M.Sc. degree (or equivalent) in image processing, computer science and/or engineering, applied mathematics, or related areas, with a strong background in linear algebra. Experience in optics and optimization are considered a plus. Candidates are also expected to have good skills in scientific programming (preferably Matlab and/or C), proficiency in English, and availability to travel within the MacSeNet network.

Applications

To apply for the position, please provide:

  1. A letter of motivation (maximum one page) explaining how your research interests, skills and experience are relevant to the position.
  2. A detailed CV, including list of publications. If any, copies of the publications most relevant to the position.
  3. Names and contact details of three referees willing to write confidential letters of recommendation
  4. Your completed eligibility form

All materials should be emailed as a single PDF file to: alessandro.foi@tut.fi with ‘PhD application MacSeNeT ESR11' in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR12 : Non-local HOSVD methods for denoising and super-resolution imaging (FINLAND) Apply by 2015-06-20

Location: Noiseless Imaging, Finland

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*:

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary:

Start Date: August to September 2015

Duration: 36 Months

Closing Date for Applications: 2015-5-1 or 2015-06-20 (if not filled)

One Early-Stage Researcher position is available. The selected candidate will study the advanced sparse approximation techniques for image denoising, deblurring, blind deblurring and super-resolution in the context of mono-modal, multi-modal, and multispectral imaging. The success of sparsity-based techniques is highly dependent on dictionaries used for image modelling. This research investigates nonlocal, data adaptive image representations to advance the state of the art in high-accuracy and high-quality imaging. Special attention will be paid to fast algorithm implementations for multiprocessor platforms.

Planned secondments

  • Scientific: Instituto de Telecomunicações, Portugal, 6 months, for training in image analysis methods.
  • Cross-Sector: Tampere University of Technology, Finland, 3 months, for research of theoretical aspects of the HOSVD-BM3D technique.

Requirements

Candidates should hold a M.Sc. degree (or equivalent) in image processing, computer science and/or engineering, applied mathematics, or related areas, with a strong background in linear algebra. Experience in optics and optimization are considered a plus. Candidates are also expected to have good skills in scientific programming (preferably Matlab and/or C), proficiency in English, and availability to travel within the MacSeNet network.

Applications

To apply for the position, please provide:

  1. A letter of motivation (maximum one page) explaining how your research interests, skills and experience are relevant to the position.
  2. A detailed CV, including list of publications. If any, copies of the publications most relevant to the position.
  3. Names and contact details of three referees willing to write confidential letters of recommendation
  4. Your completed eligibility form

All materials should be emailed as a single PDF file to: ni@noiselessimaging.com with ‘PhD application MacSeNeT ESR12' in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR13 : Audio Restoration and Inpainting (UK)

Location: CVSSP, University Of Surrey, United Kingdom

Marie Curie Annual Allowance (Pre-Employer/Employee Tax)*: €44,896

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: £29,500/£32,900 (Gross including Mobility & Family)

Start Date: July 2015

Duration: 36 Months

Closing Date for Applications: 2015-04-30

This exciting project will develop novel algorithms for audio restoration and inpainting.

Speech and music signals are often subject to localized distortions, where the intervals of distorted samples are surrounded by undistorted samples. Examples include impulsive noise or clicks, clipping, CD scratches, or packet loss in voice over IP (VoIP). A particular example is audio clipping. Here, typically either the gain at the microphone is too large for the maximum signal range that can be captured at the input, or the audio signal is amplified during processing giving a result that is too large for a subsequent representation range. These are often non-linear distortions, so nonlinear restoration methods are required. Audio inpainting was recently proposed as a framework for recovery of portions of audio data distorted due to impairments such as impulsive noise, clipping, and packet loss.

The candidate will investigate and develop new methods, incorporating more realistic signal and distortion models for complex nonlinear audio restoration, and challenging audio inpainting problems. The task can be tackled by dividing the audio signal into frames, identifying the missing portions, and then estimating the missing sections. For the particular task of audio declipping, additional information is available since the original signal is known to have a larger magnitude than the clipped version in the missing sections, and this knowledge can be used as an additional constraint on the problem. Different types of distortion and related problems will be considered, such as nonlinear "soft clipping", baseline drift, zero crossing distortion and tape saturation. The results will be evaluated using trained expert listeners and compared to existing commercial and research state-of-the-art approaches.

CVSSP is one of the major research centres of Surrey’s Department of Electronic Engineering (EE), the top ranked UK EE department in both the RAE 2008 and in the national league tables. CVSSP is the largest research centres in the UK focusing on Computer Vision, graphics and signal processing, with 120+ members comprising academic and support staff, research fellows and PhD students.

Informal enquires are welcome and should be made to Dr Wenwu Wang or Prof Mark Plumbley.

Planned secondments

  • Scientific: INRIA, France, 3 months, for training in machine learning and structured norms.
  • Cross-Sector: Cedar Audio, UK, 3 days/month, for training and applications in real-world audio restoration.

Requirements

The successful applicant is expected to have an excellent mathematical and programming background, with an MSc or equivalent related to signal processing, machine learning, and audio engineering. Programming skills in Matlab or C/C++ are required as is a high level of English (IELTS average 6.5)

Applications

Application is online via jobs.surrey.ac.uk (job ref 017715).

Please ensure your application includes

  1. your completed eligibility form.
  2. a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position.
  3. Transcripts of your undergraduate (and master) studies.
  4. certificates of qualifications (such as BSc and MSc degree certificates, and any other certificates of excellence).

Direct link to this advert (right click and copy link).

(Position Filled) ESR14 : Sound Scene Analysis (UK) Apply by 2015-04-30

Location: CVSSP, University Of Surrey, United Kingdom

Marie Curie Annual Allowance (Pre-Employer/Employee Tax)*: €44,896

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: £29,500/£32,900 (Gross including Mobility & Family)

Start Date: July 2015

Duration: 36 Months

Closing Date for Applications: 2015-04-30

This exciting project will develop new algorithms suitable for sound scene analysis applications, such as sound-based security alarm systems or home acoustic event detection.

Human listeners appear to have an innate and effortless ability to isolate, identify, and attend to one sound source while suppressing others. In this project, we would like to be able to explore computer algorithms to explore this task, to find the direction and identity of sound signals in an audio scene, and also to extract one identified sound source from a mixture. There has been relatively little research on analysis and recognition of non-speech, non-music sound scenes. An increasing interest emerges recently in this area, such as the international Multimedia Event Detection (MED) track of the TRECVID competition (since 2010) which concerns classification of audiovisual archive clips such as “baking a cake” or “birthday party”, and a data challenge on Detection and classification of acoustic scenes and events (D-CASE) which attracted 18 submissions, using a wide variety of methods from spectro-temporal modulations classified with support vector machines (SVMs), to deep learning with sparse restricted Boltzmann machines (RBMs).

The candidate will investigate new methods for sound scene analysis, building structural models based on separate analysis of “foreground” sound events and “background” sound textures. We expect to use structured sparse models and tensor models, and develop new models based on dynamic timbre of sounds inspired by physical object models. The candidate will develop a methodology for the challenging problem of audio source identification and separation. By specifically identifying sources and their characteristics, we will also expect to obtain higher quality source separation than generic audio source separation methods. Deep learning based feature representation and sound identification techniques will also be studied.

CVSSP is one of the major research centres of Surrey’s Department of Electronic Engineering (EE), the top ranked UK EE department in both the RAE 2008 and in the national league tables. CVSSP is the largest research centres in the UK focusing on Computer Vision, graphics and signal processing, with 120+ members comprising academic and support staff, research fellows and PhD students.

Informal enquires are welcome and should be made to Dr Wenwu Wang or Prof Mark Plumbley.

Planned secondments

  • Scientific: Fraunhofer IDMT, Germany, 3 months, for training in physical modelling of sound objects.
  • Cross-Sector: Audio Analytic, UK, 3 days/month, for training and applications in real-world audio event detection.

Requirements

The successful applicant is expected to have an excellent mathematical and programming background, with an MSc or equivalent related to signal processing, machine learning, and audio engineering. Programming skills in Matlab or C/C++ are required as is a high level of English (IELTS average 6.5)

Applications

Application is online via jobs.surrey.ac.uk (job ref 017815).

Please ensure your application includes

  1. your completed eligibility form.
  2. a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position.
  3. Transcripts of your undergraduate (and master) studies.
  4. certificates of qualifications (such as BSc and MSc degree certificates, and any other certificates of excellence).

Direct link to this advert (right click and copy link).

(Position Filled) ESR15 : Music source separation beyond sparse decomposition (GERMANY) Apply by 2015-06-15

Location: Fraunhofer Institute for Digital Media Technology, Ilmenau, Germany; Ph.D. and supervision at Ilmenau University of Technology

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €36,872

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €7,200 & €6,000

Approximate Country Specific Comparable salary: ca €26,000 (net, assuming single and with no children according to german tax rules)

Start Date: From June 2015

Duration: 36 Months

Closing Date for Applications: 2015-06-15

This position is about musical source modelling and separation and remixing in general. We have, for instance, multiple musical sources in a 1 channel mix, and would like to separate them. State-of-the art approaches use a time/frequency decomposition and, for instance, Non-Negative Matrix Factorization to achieve this goal.

To improve the separation and the modelling, the objectives of this position are to:

  • Explore parameters for synthesis models of music instruments
  • Use the models to find improved algorithms to separate music recordings into solo instrument and accompaniment.

Example application scenarios are a musical instrument learning software, or audio coding systems which process and render sources and instruments individually.

For information about the corresponding group at Fraunhofer IDMT see the website.

The Ph.D. supervision is with Prof. Schuller at the Ilmenau University of Technology.

Planned secondments

  • Scientific: University of Surrey, UK, 3 months, for training in audio source separation and analysis.
  • Cross-Sector: Songquito, Germany, 3 months, for training in applications to commercial music games software.

Requirements

The successful applicant is expected to have an excellent mathematical and programming background, with a valid Master’s degree with a specialisation in digital signal processing, audio processing or related fields. Programming skills in Matlab, Python or C/C++ are required.

Applications

Please ensure your application includes :

  1. Your completed eligibility form
  2. A letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position
  3. Transcripts of your undergraduate (and masters) studies.
  4. Certificates of qualifications (such as BSc and MSc degree certificates, and any other certificates of excellence).

All materials should be emailed as a single PDF file (<5 Mb) to: gerald.schuller@idmt.fraunhofer.de with ‘PhD application MacSeNet ESR15′ in the subject line.

Direct link to this advert (right click and copy link).

(Position Filled) ESR16 : Sparse models and algorithms for data on large graphs (SWITZERLAND) Apply by 2015-04-15

Location: LTS2, EPFL,Switzerland

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €49,409

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €6,000

Approximate Country Specific Comparable salary: 60,234 CHF (Three-year salary and allowances will be determined precisely by the EPFL Human Resources.)

Start Date: Feb-May 2015

Duration: 36 Months

Closing Date for Applications: 2015-04-15

Description:

One of the challenges raised by the Big Data era is that the vast majority of data generated these days is unstructured. A typical user on a social network services for instance will upload text, images or movies but will also leave a trail of items bought, places visited etc. Companies generate petabytes of unstructured data about their products or users and the rise of the connected devices (Internet of Things) adds even more versatile content. One way of bringing order to these vast datasets is to use or build relationships between items atop which large graphs can be constructed, providing the basis of discrete geometric regularity.

Objectives:

The aim of this individual project is to exploit the structure of large graphs to process information in large unstructured datasets. In particular, we will leverage the recently introduced tools of signal processing on graphs39 and focus on extending this computational framework to the Big Data scale. We will seek to formulate principles coarse to fine strategies for computing on very large graphs and study how known generative models (classes of random graphs) can be used to speed up computations.

Expected Results:

New methods for signal processing on graphs suitable for Big Data applications

Laboratory:

This is a doctoral student position open at the Laboratory of Signal Processing 2, LTS2, at EPFL, Lausanne, Switzerland, focusing on the application of compressive sensing and sparsity based methods to machine learning with applications. The LTS2 is specialized in signal/image/data processing, graph and networks analysis as well as machine learning. The lab offers a stimulating research environment with an open minded and collaborative team, state of the art IT facilities (multicore servers and GPU units). Informal enquires are welcome and should be made to Prof. Pierre Vandergheynst.

Conditions of employment:

This is a full time position as PhD researcher within the Laboratory of Signal Processing 2. The candidate will dedicate its time to research as well as be interacting with Master students and participate in teaching.

Planned secondments

  • Scientific: CTI, Greece, 3 months, for training on Big Data methods.
  • Cross-Sector: VisioSafe, Switzerland, 3 days/month for training on Big Data in video tracking.

Requirements

A candidate with an MSc degree or equivalent either in electrical engineering, computer science or applied mathematics, with a good experience in programming and scientific numerical computations. Experience with programming in one or more languages is mandatory, a good knowledge of Matlab or Python is a plus. We expect a good level of English. Ability to travel within the network is essential

Applications

To apply the candidate must register and get admitted to the EPFL doctoral school « EDIC » before 15th of April 2015.

Please ensure your application includes your completed eligibility form and a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position;

Direct link to this advert (right click and copy link).

(Position Filled) ESR17 : Towards efficient processing of 4D point clouds (SWITZERLAND) Apply by 2015-04-15

Location: LTS2, EPFL,Switzerland

Marie Sklodowska-Curie Annual Allowance (Pre-Employer/Employee Tax)*: €49,409

Marie Curie Annual Mobility Allowance & Family Allowance (Pre-Employer/Employee Tax)*: €6,000

Approximate Country Specific Comparable salary: 60,234 CHF (Three-year salary and allowances will be determined precisely by the EPFL Human Resources.)

Start Date: Feb-May 2015

Duration: 36 Months

Closing Date for Applications: 2015-04-15

Description:

The rise of cheap, high quality visual sensors and commodity depth sensors (active or passive depth cameras, stereo cameras) mean we are now able to easily construct 4D, i.e. 3D + time, accurate models of a scene. A natural support for 4D images is a data structure known as time-varying point cloud (4DPCD). These objects allow immersive visualization of photo-accurate, time-varying 3D models and might, in the near future, become as ubiquitous as standard videos.

Objectives:

In this individual project, we aim at developing models of 4DPCDs to specifically serve as the basis of future generation vision based applications. However 4DPCDs are very large objects, typically represented as time-series of 3D point clouds. To efficiently process them, one needs natural low dimension models. First we wish to develop new models for concisely describing 4DPSCD by means of dedicated feature points. This would allow to register 4DPCDs one to another, compare them and, generally, use them in recognition applications just like standard images. Second, we wish to develop dedicated structured sparse models of 4DPCDs to solve traditional problems that are very common on this type of date such as de-noising and missing data imputations. A particular emphasis will be put on making use of commodity Graphics Processing Units for efficient computation.

Expected Results:

New and efficient methods for analysis of 4D point clouds from time-varying images

Laboratory:

This is a doctoral student position open at the Laboratory of Signal Processing 2, LTS2, at EPFL, Lausanne, Switzerland, focusing on the application of compressive sensing and sparsity based methods to machine learning with applications. The LTS2 is specialized in signal/image/data processing, graph and networks analysis as well as machine learning. The lab offers a stimulating research environment with an open minded and collaborative team, state of the art IT facilities (multicore servers and GPU units). Informal enquires are welcome and should be made to Prof. Pierre Vandergheynst.

Conditions of employment:

This is a full time position as PhD researcher within the Laboratory of Signal Processing 2. The candidate will dedicate its time to research as well as be interacting with Master students and participate in teaching.

Planned secondments

  • Scientific: Instituto de Telecomunicações, Portugal, 3 months, for training in efficient methods for sequence analysis.
  • Cross-Sector: VisioSafe, Switzerland, 3 days/month for applications of 4D point clouds in video analysis and tracking.

Requirements

A candidate with an MSc degree or equivalent either in electrical engineering, computer science or applied mathematics, with a good experience in programming and scientific numerical computations. Experience with programming in one or more languages is mandatory, a good knowledge of Matlab or Python is a plus. We expect a good level of English. Ability to travel within the network is essential.

Applications

To apply the candidate must register and get admitted to the EPFL doctoral school « EDIC » before 15th of April 2015.

Please ensure your application includes your completed eligibility form and a letter of motivation including a maximum 1-page statement explaining how your research interests, skills and experience are relevant to the position;

Direct link to this advert (right click and copy link).

* Salary adjusted for host country prior to employer and employee tax being deducted, as such the gross and net amount received by the ESR will be different from that listed. Where possible the mobility allowance will be untaxed but this will depend on country tax regulations. Salaries converted into local currency are approximate at the time of writing and may change, each ESR will receive their full Euro entitlement after taxes regardless of the host currency.

MacSeNet has 17 ESRs, 15 funded by the EU's H2020 programme and 2 by the Swiss National Research funding body.

ESR1 : Tatiana Shpakova (INRIA)

Tatiana came to INRIA Paris from Moscow, Russia. She has received her M.Sc. in applied mathematics and computer science at Moscow Institute of Physics and Technology (MIPT) in 2015. She has also attended Yandex School of Data Analysis during her masters. At the moment Tatiana is a second-year Ph.D. student and she works on sub-modular-based graphical models.

ESR2 : Dmitry Babichev (INRIA)

During his high school Dmitry has participated in mathematical olympiads and has received a gold medal at the International Mathematics Olympiad (IMO) in 2008. He has received his master degree in 2014 in applied mathematics and physics at Moscow Institute of Physics and Technology. At the moment he is a second-year Ph.D student in machine learning at INRIA/ENS, Paris.

ESR3 : `Billy' Junqi Tang (Uni Edinburgh)

Billy got his Bachelor's Degree in Communication Engineering at Sichuan University, China in 2014. He started his Master's program 'Signal Processing and Communications' in the University of Edinburgh in September, 2014 and finishes his study as the class medal winner. He did his master's thesis 'The Non-uniform Fast Fourier Transform in Computed Tomography' with his supervisor Prof. Mike Davies. Billy is currently developing fast structure-exploiting optimization algorithms for large scale inverse problems, signal/image processing and machine learning as a PhD student at the University of Edinburgh, under the supervision of Prof. Mike Davies.

ESR4 : Arnold Benjamin (Uni Edinburgh)

Arnold Julian Vinoj Benjamin hails from Chennai (Madras), a city in South India. He completed his Bachelor's degree in Electronics and Communication Engineering from Anna University, Chennai in 2012. His great interest in image/signal processing helped him to get accepted into a Masters Research program in Biomedical Engineering at National University of Singapore (NUS) in Aug 2012. He was part of the Biophotonics and Bioimaging Laboratory in the Department of Biomedical Engineering at NUS.

His master’s thesis is titled, ‘Real-Time Monitoring of Probe-Tissue Pressure Effects on In Vivo Skin Optical Spectroscopy Using a Pressure Sensitive Fiberoptic Probe’. Its main application was real-time cancer diagnosis using optical spectroscopy. It involved the real-time evaluation of probe pressure induced spectral variability on various optical techniques like Raman, Fluorescence and Diffuse Reflectance Spectroscopy. It included the design and development of an integrated novel real-time pressure sensitive probe for in vivo tissue measurements. This comprehensive study on probe pressure induced optical spectral variability aimed to increase the diagnostic capability of various optical spectroscopy techniques by characterizing the effect of exerted probe pressure on the acquired spectra.

He started his PhD at the University of Edinburgh in September 2015 and his project is titled, `Next generation compressed sensing techniques for quantitative MRI'. The major objective of the research project is to accelerate Magnetic Resonance Imaging (MRI) data acquisition by using a combination of compressed sensing (CS) and parallel imaging (PI) in order to reduce total scan time for patients. Specifically, it looks at implementing a fast scanning scheme that utilizes highly sub-sampled data to reconstruct good quality magnetic resonance (MR) images that are deemed clinically acceptable by radiologists.

ESR5 : Dhritiman Das (TUM)

Dhritiman was born and raised in New Delhi, India, and completed his Bachelor of Engineering from Manipal University, India. This was followed by a Master of Science in Bioengineering from the Arizona State University. His Masters work involved designing an optical-flow based algorithm to improve the resolution of Particle Image Velocimetry (PIV) images (useful for visualizing fluid flow in aneurysm models). He eventually gave this work a nice-sounding, scientific title - "A minimal divergence interpolator for fluid flow velocity images".

He has also worked in the Imaging and Computer Vision group at Siemens Corporate Research in projects related to optical flow and motion magnification for X-ray images. Currently he is pursuing his PhD in Computer Science in the domain of machine learning and computer vision. His work involves developing sparsity-based and data-driven methods for advanced reconstruction and post-processing of MR Spectroscopy Images (MRSI). His research interests include image processing, computer vision and machine learning with applications to medical imaging.

ESR6 : Cagdas Ulas (TUM)

Cagdas Ulas received his Bachelor and Master degree both in Electronics Engineering, 2011 and 2013, respectively, from Sabanci University, Turkey. During his Master study, he focused on integrating statistical language models into P300-based brain-computer interfaces. After successfully obtaining his Master degree in July 2013, he immediately joined the R\&D department of a Turkish bank as a research engineer. During one and half year period in industry, he worked on several projects that mainly involve different applications of machine learning, computer vision, text classification and natural language processing. On April 2015, he moved to Germany and joined Technical University of Munich (TUM) as a PhD candidate. Currently he is pursuing his PhD in compressed-sensing based reconstruction schemes applied on perfusion magnetic resonance (MR) image sequences, under the MacSeNet project at TUM. His research interests include inverse problems in image processing, machine learning, and optimization methods applied on medical data

ESR7 : Manuel Morante Moreno (CTI)

Manuel Morante Moreno received his bachelor of Science degree in Physics in 2014 from the University of Granada (Spain). After that, he continued to a Master degree in Nanotechnology: Physics and applications. In 2015 he finished his Master's thesis: Study of Methods for the Analysis of Transients at the University of Granada (Spain). Currently, he is working towards the PhD degree at the University of Athens (Greece), within the MacSeNet project, focusing on the study of functional MRI via distributed dictionary learning.

ESR8 : Christos Chatzichristos (CTI)

Christos Chatzichristos was born in Thessaloniki, Greece, in July 1986. He graduated from the Aristotle University of Thessaloniki with the Diploma (MEng) of Electrical and Computer Engineering. In his Diploma thesis he was introduced into bioinformatics, through development of a Web based application for the correlation of phylogenetic profiles with metabolic pathways in order to extract evolutionary motives. This motivated him to continue his studies in Biomedical engineering. Hence, after having complete his military service, he moved to Katholieke Universiteit Leuven (KU Leuven) for a Masters of Biomedical Engineering.

During his internship, in Icometrix NV, Leuven, and his Masters thesis research on "The effect of the fiber response kernel on fiber tracking in diffusion MRI", he worked with different modalities of Magnetic Resonance Imaging (MRI). Currently he is pursuing his PhD studies in "Functional neuroimaging data characterization via tensor representations", within the MacSeNet project at the National and Kapodistrian University of Athens, Greece. His research interests include (Biomedical) Signal Processing, (Medical) Image Analysis and Machine Learning.

ESR9 : Joshin Krishnan (IT)

Joshin P. Krishnan is currently pursuing PhD degree at Instituto de Telecomunicacoes, Instituto Superior Tecnico, Lisbon, Portugal. He is from India where he has done his schooling, graduation and masters degree. His Bachelors degree is in Electronics and Communication from College of Engineering, Trivandrum. Afterwards, he worked as VLSI Design Engineer at Coreel Technologies, Bangalore for two years. Later he pursued his masters in Telecommunication from Indian Institute of Science, Bangalore. He also has 1.5 years of experience as Assistant Professor at PESIT South Campus, Bangalore after his post graduation. His present PhD thesis is in the topic Interferometric Phase Estimation, supervised by Prof. Jose M. Bioucas-Dias and Prof. Mario Figueiredo.

ESR10 : Marina Ljubenovic (IT)

Marina Ljubenovic, born Petrovic, received a Bachelor with honours in Electrical and Computer Engineering at the Faculty of Technical Science, University of Novi Sad, Serbia in 2010. At the same faculty, Marina received the Master's degree in Electrical and Computer Engineering in 2015. During both, bachelor and master studies, she was part of the Telecommunication and Signal Processing Group. The title of her master thesis was ``Deformable registration of Dual Energy X-ray image``. Between April 2013 and August 2015, Marina was working as Embedded engineer in Serbian company called MikroElektronika. Currently, she is pursuing a PhD degree in Instituto de Telecomunicacoes, Instituto Superior Tecnico, Lisbon, Portugal. She is working on non-local patch-based and dictionary-based methods for blind image deblurring, supervised by Prof. Mario A.T. Figueiredo and by Prof. Jose M. Bioucas-Dias.

ESR11 : Nasser Eslahi (TUT)

Nasser Eslahi received the B.S. and M.S. degrees in electrical engineering from Imam Khomeini International University and Babol University of Technology, Iran, in 2012 and 2015, respectively. He is currently working toward his Ph.D. degree in the Department of Signal Processing, Tampere University of Technology, Tampere, Finland. His research interests include image and video processing, statistical signal processing, sparse representation/approximation, compressive sensing, image inverse problems and convex optimization.

ESR12 : Cristóvão Cruz (NI)

Cristóvão André Antunes Cruz was born in 1989, in Braga, Portugal. He completed his masters degree in Electronic and Telecommunications Engineering at University of Aveiro in 2014. During his studies, he was a member of the University of Aveiro Linux Group (GLUA) and pursued several different research subjects, including reconfigurable computing based networks, artificial vision for mobile robots and vehicular communications. For his master thesis he helped develop a wireless communications device for vehicular networks, in a industry driven project. Later on, he worked for one year at Tampere University of Technology tackling video transmission over random loss channels. He is currently working at Noiseless Imaging Oy and enrolled on the doctoral programme of Computing and Electrical Engineering at Tampere University of Technology. His research interests include image restoration, high-order single value decomposition and sparse representation.

ESR13 : Lucas Rencker (SURREY)

Lucas Rencker graduated from Ecole Centrale Marseille (Marseille, France) in Electronic Engineering and Signal Processing. He did his masters thesis at Audiolabs Erlangen, on the topic of impulsive noise reduction for speech enhancement. Currently, he is pursuing a PhD on audio restoration, at the Centre for Vision, Speech and Signal Processing (CVSSP), University of Surrey. His research interests include sparse representations and compressed sensing, with applications to audio denoising and inpainting.

ESR14 : Iwona Sobieraj (SURREY)

Iwona Sobieraj received her master degree in Telecommunications from Warsaw University of Technology with a thesis on audio watermarking accomplished at Fraunhofer IIS in Germany, where she stayed for 14 months, first as an intern, then as an employee. After completing her degree she worked for 3 years on machine translation in Samsung Electronics. Currently she is interested in machine learning for audio processing, focusing on environmental audio. More specifically, her PhD topic is Sound Event Detection in real life environments.

ESR15 : Stylianos Ioannis Mimilakis (F-IDMT)

Stylianos Ioannis Mimilakis received a Master of Science degree in Sound & Music Computing from Pompeu Fabra University and a Bachelor of Engineering in Sound & Music Instruments Technologies from Higher Technological Education Institute of Ionian Islands. Currently he is pursuing his PhD in audio signal processing for music source separation, under the MacSeNet project at Fraunhofer IDMT. His research interests include, inverse problems in audio processing and synthesis, and deep learning.

ESR16 : Volodymyr Miz (EPFL)

Volodymyr Miz received the M.S. degree in computer engineering from National University of Radio Electronics (NURE), Kharkov, Ukraine, in 2013. The title of his Master theses was ``Motion Detection in Constrained Environment''. Volodymyr investigated motion detection algorithms and figured out that the combination of algorithms that process images on different levels (pixel, region and frame) shows the best result in difficult environmental conditions. As a result, during an internship in the USA based telecommunication company EchoStar, he developed business logic for the motion detection surveillance system and completed architecture design of the system. Successful internship allowed Volodymyr to obtain a permanent Software Engineer position in the company where he stayed for more than 3 years till March 2016. His research interests include graph theory, data mining and time-series signal processing.

ESR17 : Youngjoo Seo (EPFL)

Youngjoo is currently a Ph.D candidate at LTS2 (Signal processing laboratory2) of EPFL. His research interest mainly focuses on Deep Neural Networks on graph structured data. Especially, He is focused on developing a novel Recurrent Neural Networks (RNNs) model on Spatio-temporal structured data which considering time and space together in one unified network. Before joining EPFL, He finished his Masters at Korea Advanced Institute of Science and Technology (KAIST) on computer vision such as image segmentation and object detection. Also, he worked start-up company named solidware. As a machine learning researcher, He developed core algorithm to predict credit score on an individual person using deep regression model.

Full Partners


Centre For Vision Speech and Signal Processing
University of Surrey
www.surrey.ac.uk/CVSSP
Mark Plumbley
Wenwu Wang
Institut National de Recherche en Informatique et en Automatique
www.inria.fr / www.di.ens.fr
Francis Bach
Institute for Digital Communications
The University of Edinburgh
www.eng.ed.ac.uk/research/institutes/idcom
Mike Davies
Technical University Munich
campar.in.tum.de
Bjoern Menze
Computer Technology Institute
www.cti.gr/en/
Sergios Theodoridis
Instituto de Telecomunicações (IT)
www.it.pt
Jose Bioucas Dias
Mario Figueiredo
Tampere University of Technology
www.tut.fi
Alessandro Foi
Noiseless Imaging
www.noiselessimaging.com
Fraunhofer-Gesellschaft zur Foerderung der Angewandten Forschung E.V.
www.idmt.fraunhofer.de
LTS2
Ecole Polytechnique Fédérale de Lausanne
lts2www.epfl.ch
Pierre Vandergheynst
Xavier Bresson
Visiosafe S.A.
www.visiosafe.com

Associate Partners

General Electric
www.ge.com
Bioiatriki
bioiatriki.gr
Audio Analytic
www.audioanalytic.com
Cedar Audio Ltd.
www.cedar-audio.com
Songquito
www.songs2see.com

Peer Reviewed Publications

2017
[21] M.M. Moreno, Y. Kopsinis, E. Kofidis, C. Chatzichristos, Theodoridis S., "Assisted Dictionary Learning for fMRI Data Analysis", In IEEE Int. Conf. Acoust., Speech, and Signal Process. (ICASSP), New Orleans, USA, 2017. [bib]
[20] C. Chatzichristos, E. Kofidis, S. Theodoridis, "PARAFAC2 and its Block Term Decomoposition Analog for Blind fMRI Source unmixing", In European Signal Processing Conference (EUSIPCO), Kos, Greece, 2017. [bib]
[19] C. Chatzichristos, E. Kofidis, Y. Kopsinis, M. Morante Moreno, S. Theodoridis, "Higher-Order Block Term Decomposition for Spatially Folded fMRI Data", In Int. Conf. on Latent Variable Analysis, and Signal Separation. (LVA ICA), Grenoble, France, 2017. [bib]
2016
[18] Tatiana Shpakova, Francis Bach, "Parameter Learning for Log-supermodular Distributions", In Advances in Neural Information Processing Systems, Barcelona, Spain, 2016. [bib]
[17] A. Liutkus S.I. Mimilakis D. Fitzgerald B. Pardo Z. Rafii, "Whitepaper: A Review of the Research on the Separation of Lead and Accompaniment in Music", In IEEE/ACM Transactions on Audio, Speech and Language Processing, 2016. [bib]
[16] S.I. Mimilakis, K. Drossos, T. Virtanen, G. Schuller, "Deep Neural Networks for Dynamic Range Compression in Mastering Applications", In Audio Engineering Society Convention 140, Paris, France, 2016. [bib] [pdf]
[15] S.I. Mimilakis, E. Cano, J. Abeßer, G. Schuller, "New Sonorities for Jazz Recordings: Separation and Mixing using Deep Neural Networks", In Audio Engineering Society Workshop on Intelligent Music Production 140, London, UK, 2016. [bib]
[14] C. Bonsel, J. Abeßer, S. Grollmisch, S.I. Mimilakis, "Automatic Best Take Detection for Electric Guitar and Vocal Studio Recordings", In Audio Engineering Society Workshop on Intelligent Music Production 140, London, UK, 2016. [bib]
[13] Cagdas Ulas, Pedro A Gomez, Felix Krahmer, Jonathan I Sperl, Marion I Menzel, Bjoern H Menze, "Robust Reconstruction of Accelerated Perfusion MRI Using Local and Nonlocal Constraints", In Proceedings of MICCAI Workshop on Reconstruction and Analysis of Moving Body Organs, Athens,Greece, 2016. [bib]
[12] Cagdas Ulas, Pedro A Gomez, Jonathan I Sperl, Christine Preibisch, Bjoern H Menze, "Spatio-temporal MRI Reconstruction by Enforcing Local and Global Regularity via Dynamic Total Variation and Nuclear Norm Minimization", In Proceedings of International Symposium on Biomedical Imaging (ISBI), Prague, Czech Republic, 2016. [bib]
[11] Pedro A Gómez, Miguel Molina-Romero, Cagdas Ulas, Guido Bounincontri, Jonathan I Sperl, Jones Derek K, Marion I Menzel, Bjoern H Menze, "Simultaneous Parameter Mapping, Modality Synthesis, and Anatomical Labeling of the Brain with MR Fingerprinting", In Proceedings of the 19th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), Athens, Greece, 2016. [bib]
[10] Pedro A Gómez, Guido Bounincontri, Miguel Molina-Romero, Cagdas Ulas, Jonathan I Sperl, Marion I Menzel, Bjoern H Menze, "3D Magnetic Resonance Fingerprinting with a Clustered Spatiotemporal Dictionary", In Proc Intl Soc Mag Reson Med, Singapore, 2016. [bib]
[9] Dhritiman Das, Eduardo Coello, Rolf F Schulte, Bjoern H Menze, "Spatially Adaptive Spectral Denoising for MR Spectroscopic Imaging using Frequency-Phase Non-Local Means", In International Conference on Medical Image Computing and Computer Assisted Interventions (MICCAI), Athens, Greece, 2016. [bib]
[8] Dhritiman Das, Eduardo Coello, Rolf F Schulte, Bjoern H Menze, "Spectral denoising for MR Spectroscopic Imaging using Non-Local Means", In ISMRM Workshop on MR Spectroscopy, Lake Konstanz, 2016. [bib]
[7] Arnold J. V. Benjamin, Wahija Bano, Ian Marshall, Mike E. Davies, "Optimizing calibration kernels and sampling pattern for ESPIRiT based compressed sensing implementation in 3D MRI", In Proceedings of the 22nd Annual Scientific Meeting of the British Chapter of the ISMRM, 2016. [bib]
[6] Joonas Iisakki Multanen, Heikki Kultala, Matias Koskela, Timo Viitanen, Pekka Jäskeläinen, Jarmo Takala, Aram Danielyan, Cristóvão Cruz, "OpenCL Programmable Exposed Datapath High Performance Low-Power Image Signal Processor", In IEEE Nordic Circuits and Systems Conference, 2016. [bib] [pdf]
[5] E. Kofidis, C. Chatzichristos, A. L.F. Almeida, "Joint Channel Estimation/ Data Detection in MIMO-FBMC/OQAM Systems - A Tensor Based Approach", In Workshop on Tensor Decompositions and Applications (TDA-2016), Leuven, Belgium, 2016. [bib] [pdf]
[4] Iwona Sobieraj, Mark D. Plumbley, "Coupled Sparse NMF vs. Random Forest Classification for Real Life Acoustic Event Detection", In Proceedings of the Detection and Classification of Acoustic Scenes and Events 2016 Workshop (DCASE2016), pp. 90-94, 2016. [bib]
[3] E. Cano J. Abeßer G. Schuller S.I. Mimilakis, "New Sonorities for Jazz Recordings: Separation and Mixing using Deep Neural Networks", In 2nd AES Workshop on Intelligent Music Production, 2016. [bib]
[2] Junqi Tang, Mohammad Golbabaee, Mike Davies, "Gradient Projection Iterative Sketch for Large Scale Constrained Least-Sqaures", In ArXiv e-prints, 2016. [bib] [pdf]
2015
[1] Pedro A Gómez, Cagdas Ulas, Jonathan I Sperl, Tim Sprenger, Miguel Molina-Romero, Marion I Menzel, Bjoern H Menze, "Learning a Spatiotemporal Dictionary for Magnetic Resonance Fingerprinting with Compressed Sensing", In Proceedings of MICCAI Workshop on Patch-based Techniques in Medical Imaging, Munich, Germany, 2015. [bib]

Powered by bibtexbrowser