Computational vision is one of the most challenging research domains in engineering sciences. The aim is to reproduce human visual perception through intelligent processing of visual data. The application domains span from computer aided diagnosis to industrial automation & robotics. The most common mathematical formulation to address such a challenge is through mathematical modeling. In such a context, first the solution of the desired vision task is expressed in the form of a parameterized mathematical model. Given such a model, the next task consists of associating the model parameters with the available observations, which is often called the model-to-data association. The aim of this task is to determine the impact of a parameter choice to the observations and eventually maximize/minimize the adequacy of these parameters with the visual observations. In simple words, the better the solution is, the better it will be able to express and fit the data. This is often achieved through the definition of an objective function on the parametric space of the model. Last, but not least given the definition of the objective function, visual perception is addressed through its optimization with respect to the model parameters. To summarize, computation visual perception involves three aspects, a task-specific definition of a parametric model, a data-specific association of this model with the available observations and last the optimization of the model parameters given the objective and the observations.

Such a chain processing inherits important shortcomings. The curse of dimensionality is often used to express the importance of the model complexity. In simple words, the higher the complexity of the model is, the better its expressive power will be with counter effect the increase of the difficulty of the inference process. Non-linearity is another issue to be addressed which simply states that the association between the model and the data is a (highly) non-linear function and therefore direct inference is almost infeasible. The impact of this aspect is enforced from the curse of non-convexity that characterizes the objective function. Often it lives in high-dimensional spaces and is ill posed making exact inference problematic (in many cases not possible) and computationally expensive. Last, but not least modularity and scalability is another important concern to be addressed in the context of computational vision. The use of task-specific modeling and algorithmic solutions make their portability infeasible and therefore transfer of knowledge from one task to another is not straightforward while the methods do not always scale well with respect either to the dimensionality of the representation or the data.

GALEN aims at proposing innovative techniques towards automatic structuring, interpretation and longitudinal modeling of visual data. In order to address these fundamental problems of computational perception, GALEN investigates the use of discrete models of varying complexity. These methods exhibit an important number of strengths such as their ability to be modular with respect to the input measurements (clinical data), the nature of the model (certain constraints are imposed from computational perspective in terms of the level of interactions), and the model-to-data association while being computational efficient.

A general framework for the fundamental problems of image segmentation, object recognition and scene analysis is the interpretation of an image in terms of a set of symbols and relations among them. Abstractly stated, image interpretation amounts to mapping an observed image, *optimally explain the underlying image*, as measured by a scoring function

Applying this framework requires (a) identifying which symbols and relations to use (b) learning a scoring function

Applying this framework requires (a) identifying which symbols and relations to use for image and object representation (b) learning a scoring function s from training data and (c) optimizing over Y in Eq. . One of the main themes of our work is the development of methods that jointly address (a,b,c) in a shape-grouping framework in order to reliably extract, describe, model and detect shape information from natural and medical images. A principal motivation for using a shape-based framework is the understanding that shape- and more generally, grouping- based representations can go all the way from image features to objects. Regarding aspect (a), image representation, we cater for the extraction of image features that respect the shape properties of image structures. Such features are typically constructed to be purely geometric (e.g. boundaries, symmetry axes, image segments), or appearance-based, such as image descriptors. The use of machine learning has been shown to facilitate the robust and efficient extraction of such features, while the grouping of local evidence is known to be necessary to disambiguate the potentially noisy local measurements. In our research we have worked on improving feature extraction, proposing novel blends of invariant geometric- and appearance- based features, as well as grouping algorithms that allow for the efficient construction of optimal assemblies of local features.

Regarding aspect (b) we have worked on learning scoring functions for detection with deformable models that can exploit the developed low-level representations, while also being amenable to efficient optimization. Our works in this direction build on the graph-based framework to construct models that reflect the shape properties of the structure being modeled. We have used discriminative learning to exploit boundary- and symmetry-based representations for the construction of hierarchical models for shape detection, while for medical images we have developed methods for the end-to-end discriminative training of deformable contour models that combine low-level descriptors with contour-based organ boundary representations.

Regarding aspect (c) we have developed algorithms which implement top-down/bottom-up computation both in deterministic and stochastic optimization. The main idea is that ‘bottom-up’, image-based guidance is necessary for efficient detection, while ‘top-down’, object-based knowledge can disambiguate and help reliably interpret a given image; a combination of both modes of operation is necessary to combine accuracy with efficiency. In particular we have developed novel techniques for object detection that employ combinatorial optimization tools (A* and Branch-and-Bound) to tame the combinatorial complexity, achieving a best-case performance that is logarithmic in the number of pixels.

In the long run we aim at scaling up shape-based methods to 3D detection and pose estimation and large-scale object detection. One aspect which seems central to this is the development of appropriate mid-level representations. This is a problem that has received increased interest lately in the 2D case and is relatively mature, but in 3D it has been pursued primarily through ad-hoc schemes. We anticipate that questions pertaining to part sharing in 3D will be addressed most successfully by relying on explicit 3D representations. On the one hand depth sensors, such as Microsoft’s Kinect, are now cheap enough to bring surface modeling and matching into the mainstream of computer vision - so these advances may be directly exploitable at test time for detection. On the other hand, even if we do not use depth information at test time, having 3D information can simplify the modeling task during training. In on-going work with collaborators we have started exploring combinations of such aspects, namely (i) the use of surface analysis tools to match surfaces from depth sensors (ii) using branch-and-bound for efficient inference in 3D space and (iii) groupwise-registration to build statistical 3D surface models. In the coming years we intend to pursue a tighter integration of these different directions for scalable 3D object recognition.

The foundation of statistical inference is to learn a function that minimizes the expected loss of a prediction with respect to some unknown distribution

where

A key problem is that the distribution

which makes an implicit assumption that the training samples

where

Equation () is very well studied in classical statistics for the case that the output,

Many tasks in artificial intelligence are solved by building a model whose parameters encode the prior domain knowledge and the likelihood of the observed data.
In order to use such models in practice, we need to estimate its parameters automatically using training data. The most prevalent paradigm of parameter estimation is supervised learning, which requires the collection of the inputs

In order to address the deficiencies of supervised learning, researchers have started to focus on the problem of parameter estimation with data that contains hidden
variables. The hidden variables model the missing information in the annotations. Obtaining such data is practically more feasible: image-level labels (`contains car',`does not contain person') instead of tight bounding boxes; partial segmentation of medical images. Formally, the parameters **w** of the model are learned by minimizing the following objective:

Here,

Previous attempts at minimizing the above objective function treat all the training samples equally. This is in stark contrast to how a child learns: first focus on easy samples (`learn to add two natural numbers') before moving on to more complex samples (`learn to add two complex numbers'). In our work, we capture this intuition using a novel, iterative algorithm called self-paced learning (spl). At an iteration

Here, samples with

Here, **w**. The use of a more general loss function will allow us to better exploit the freely available data with missing information. For example, consider the case where

A wide variety of tasks in medical image analysis can be formulated as discrete labeling problems. In very simple terms, a discrete optimization problem can be stated as follows: we are given a discrete set of variables *singleton* potential function *pairwise* potential function

Our goal is then to choose a labeling which will allow us to pay the smallest total price. In other words, based on what we have mentioned above, we want to choose a labeling that minimizes the sum of all the MRF potentials, or equivalently the MRF energy. This amounts to solving the following optimization problem:

The use of such a model can describe a number of challenging problems in medical image analysis. However these simplistic models can only account for simple interactions between variables, a rather constrained scenario for high-level medical imaging perception tasks. One can augment the expression power of this model through higher order interactions between variables, or a number of cliques

where

The use of contrast enhanced imaging is investigated in collaboration with the Montpellier University Hospital towards better understanding of low-gliomas positioning, automatic tumor segmentation/identification and longitudinal (tumor) growth modeling. Furthermore, in collaboration with the Neurospin center of CEA and the Brookhaven National Laboratory at StonyBrook University we investigate the use of machine learning methods towards automatic interpretation of functional magnetic resonance imaging between cocaine addicted and normal subjects. Last, but not least in collaboration with the Georges Pompidou European Hospital an effort toward understanding tumor perfusion process through comportemental models is carried out with emphasis given on elastic organs.

The use of CT and MR imaging for cancer guidance treatment in collaboration with the Gustave Roussy Institute of Oncology. The aim is to provide tools for automatic dose estimation as well as off-line and online positioning guidance through deformable fusion between imaging data prior to each session and the ones used for scheduling/planning and dose estimation. The same concept will be explored in collaboration with the Saint-Antoine University Hospital towards image-driven surgery guidance through 2D to 3D registration between interventional and pre-operative annotated data.

Retinal images–also known as fundus images or retinographies–are projective color im- ages of the inner surface of the human eye. In collaboration with Pladema Institute, UNCPBA, Argentina, we are developing a suite of software tools for automatic analysis of retinal images driven by statistical learning approaches.

deformable image and volume registration,
is a deformable registration platform in C++ for the medical imaging community (publicly available at http://

Scale-Invariant Descriptor, Scale-Invariant Heat Kernel Signatures
Disd (publicly available at http://

Average precision optimization, high-order information, ranking
The software (publicly available at http://

branch-and-bound, parts detection, segmentation,
Dpms implements branch-and-bound object detection, cutting down the complexity of detection from linear in the number of pixels to logarithmic (publicly available at http://

discrete optimization, Markov random field, duality, graph cuts,
FastPD is an optimization platform in C++ for the computer vision and medical imaging community (publicly available at http://

procedural modeling, image-based building reconstruction, shape grammars
GraPeS is a generic image parsing library based on re-inforcement learning (publicly available at http://

Scale-Invariant Descriptor, Scale-Invariant Heat Kernel Signatures
Lbsd (publicly available at http://

Texture, modulation, generative models, segmentation,
TeXMeG is a front-end for texture analysis and edge detection platform in Matlab that relies on Gabor filtering and image demodulation (publicly available at http://

Handbook of Biomedical Imaging: Methodologies and Clinical Research - co-edited from Nikos Paragios, James Duncan and Nicholas Ayache - has been published from Springer Publishing house.

Nikos Paragios was admitted as a senior fellow at the Insitut Universitaire de France and has been awarded an IBM Faculty award. He has also been one of the plenary invited lecturers at the IARP International Conference in Pattern Recognition (ICPR'2015, Stockholm).

**Paticipants:** M. Pawan Kumar

Metric labeling is an important special case of energy minimizaton in Markov random fields. While the best known polynomial-time algorithm for the problem is the linear programming (LP) relaxation, in practice it is slow to solve it. In , we introduced a new family of efficient move-making algorithms for metric labeling. These algorithms mimic the rounding procedues used for converting a fractional LP solution to a feasible integral solution. Our algorithms provide a matching theoretical guarantee to the LP relaxation, while requiring significantly less computational time.

**Paticipants:** Puneet Kumar Dokania, Aseem Behl, Pritish Mohapatra, C.V. Jawahar, M. Pawan Kumar

Average precision (AP) is one of the most commonly used measures for ranking. However, due to the inefficiency of optimizing it during learning, a common approach is to use surrogate loss functions such as 0-1 loss. In , we proposed a new optimization algorithm for AP-SVM that allows training in a similar time to binary SVM. In , we extended the AP-SVM framework to score the samples according to high-order information, as opposed to simple first-order information used in prior work. Finally, in , we proposed a novel latent AP-SVM formulation that allows learning from weakly supervised datasets. The advantage of learning with high order and missing information is demonstrated on challenging computer vision problems such as action classification and object detection using standard benchmark datasets.

**Paticipants:** Haithem Boussaid, Iasonas Kokkinos and Nikos Paragios

Deformable Contour Models (DCMs) are a main workhorse for medical image analysis - but are not commonly studied from a machine learning perspective. In , we haved proposed an integrated machine learning and optimization framework to deploy DCMs in medical image analysis.

Our technical contributions are two-fold: firstly, we use an efficient decomposition-coordination algorithm to solve the optimization problems resulting from Loopy DCMs, by means of the Alternating Direction Method of Multipliers (ADMM); this yields substantially faster convergence than plain Dual Decomposition-based methods.

Secondly, we use structured prediction to exploit loss functions that better reflect the performance criteria used in medical image segmentation. By using the mean contour distance (MCD) as a structured loss during train- ing, we obtain clear test-time performance gains.

We demonstrate the merits of exact and efficient inference with rich, structured models in a large X-Ray image segmentation benchmark, where we obtain systematic improvements over the current state-of-the-art.

**Paticipants:** Iasonas Kokkinos, Stavros Tsogkas, Eduard Trulls, Pierre-Andre Savalle, George Papandreou.

In and we have worked on improving the classification accuracy of Deformable Part Models (DPMs) for object detection in two distinct manners. Firstly, in we propose a technique to combine bottom-up segmentation, coming in the form of SLIC superpixels, with sliding window DPM detectors. The merit of our approach lies in ‘cleaning up’ the low- level features by exploiting the spatial support indicated by segmentation. - tion, for both the root and part filters of DPMs. We use these masks to construct enhanced, background- invariant features to train DPMs. We test our approach on the PASCAL VOC 2007, outperforming the standard DPM in 17 out of 20 classes, yielding an average increase of 1.7AP. Additionally, we demonstrate the robustness of this ap- proach, extending it to dense SIFT descriptors for large dis- placement optical flow.

Secondly, in we have explored the potential of convolutional neural networks as feature extractors for detection with DPMs. In particular, we substitute the Histogram-of-Gradient features of DPMs with Convolutional Neural Network (CNN) features, and demonstrate that we thereby obtain a substantial boost in performance (+14.5 mAP) when compared to the baseline HOG-based models. Some more recent extensions to this work are included in where we explore the potential of explicit scale and aspect ratio search in the context of sliding window detection with CNNs.

**Paticipants:** Iasonas Kokkinos, Matthew Blaschko, Stavros Tsogkas, Andrea Vedaldi, Mircea Cimpol, Subhransu Maji, Ross Girshick, Juho Kannala, Esa Rahtu, David Weiss, Ben Taskar, Karen Simonyan.

In and we explore methods for the fine-grained understanding of objects and textures, respectively.

**Paticipants:** Matthew Blaschko

**Paticipants:** Chaohui Wang, Dimitris Samaras, Nikos Paragios

**Paticipants:** Nikos Paragios

Grammar-like representations are powerful modeling and inference tools in computational vision. In a novel approach towards automatic inference of typology specific building grammars has been introduced. The central idea was to consider that such grammars could be derived through a bottom up approach of common sub-tree reasoning of derivation trees determined through parsing using elementary shape (binary split) grammars. Such an approach performs common subtree reduction within the entire training set and identifies meta-rules (corresponding to the same subtrees) which are then clustered together towards producing a compact, typology specific grammar. Promising results both in terms of grammar compactness as well as in terms of inference demonstrated the potentials of the method that could be used beyond the considered scoped.

**Paticipants:** Matthew Blaschko, José Ignacio Orlando

**Paticipants:** Sarah Parisot, Deepak Chittajallu, Ioannis Kakadiaris, Nikos Paragios

In we revisited explicit contour-evolution segmentation methods driven from a graph-based shape prior. Prior knowledge through geometric constraints has been encoded to the model within pair-wise interactions between control points. The segmentation process was driven through an objective function seeking to move the control points towards image locations optimizing the expected visual properties of the organ while satisfying the prior geometric constraints being learned at training. In we have proposed a mathematical formalism for automatic tumor segmentation which was taking advantage of conventional segmentation likelihoods and atlas-based segmentation methods. The central idea was to jointly deform and segment an atlas such that the tumor likelihoods are maximized once projected to the targeted image while relaxing the registration constraints in this area. Furthermore we have endowed to this framework explicit estimation of uncertainties allowing the dynamic sampling of the graph structure resulting on significant speed up of the process while producing quantitative means for the interpretation of the final result.

**Paticipants:** Stavros Alchatzidis, Aristeidis Sotiras, Nikos Paragios

**Paticipants:** Nikos Komodakis, Bo Xiang, Nikos Paragios

**Microsoft Research, Cambridge, UK**: Large Scale Diverse Learning for Structured Output Prediction [Ph.D. thesis D. Bouchacourt]

**General Electrric HealthCare, Buc, FR**: Patient-Specific Optimization of Computed Tomography Acquisition Protocols [Ph.D. thesis H. Pasquier]

Program: DIGITEO (Chair)

Project acronym: SubSample

Project title: Identification and prediction of Salient brain States through probabilistic structure learning towards fusion of imaging and genomic date

Duration: 01/2012-12/2015

Coordinator: ECP - FR

Program: DIGITEO (OMTE)

Project acronym: Curator

Project title: Real-time 2D/3D Deformable Fusion Towards Computer Assisted Surgery

Duration: 01/2013-01/2015

Coordinator: ECP - FR

Program: DIGITEO

Project acronym: SOPRANO

Project title: Structured Output Prediction on Large Scale Neuroscience Data

Duration: 3/2013-3/2016

Coordinator: Ecole Centrale Paris - FR

Program: MEDICEN

Project acronym: ADOC

Project title: ADOC – Diagnostic peropératoire numérique en chirurgie du cancer

Duration: 11/2011-09/2015

Coordinator: LLTECH - FR

Program: ANR Blanc International

Project acronym: ADAMANTIUS

Project title: Automatic Detection And characterization of residual Masses in pAtients with lymphomas through fusioN of whole-body diffusion-weighTed mrI on 3T and 18F-flUorodeoxyglucoSe pet/ct

Duration: 9/2012-8/2015

Coordinator: CHU Henri Mondor - FR

Program: ANR JCJC

Project acronym: HICORE

Project title: HIerarchical COmpositional REpresentations for Computer Vision

Duration: 10/2010-9/2014

Coordinator: ECP - FR

Program: ANR JCJC

Project acronym: LearnCost

Project title: Learning Model Constraints for Structured Prediction

Duration: 2014-2018

Coordinator: Inria Saclay - FR

Program: ITMOs Cancer & Technologies pour la santé d'Aviesan / INCa

Project acronym: CURATOR

Project title: Slice-to-Image Deformable Registration towards Image-based Surgery Navigation & Guidance

Duration: 12/2013-11/2015

Coordinator: ECP - FR

Type: FP7

Instrument: European Research Council

Duration: September 2011 - August 2016

Coordinator: Nikos Paragios

Partner: Ecole Centrale de Paris (FR)

Inria contact: Nikos Paragios

Type: FP7

Defi: Cognitive Systems and Robotics

Instrument: Specific Targeted Research Project

Objectif: Cognitive Systems and Robotics

Duration: February 2013 - January 2016

Coordinator: Angelika Peer

Partner: University of Bristol (UK)

Inria contact: Iasonas Kokkinos

Type: H2020

Defi: Cognitive Systems and Robotics

Instrument: Specific Targeted Research Project

Objectif: Cognitive Systems and Robotics

Duration: March 2015 - February 2018

Coordinator: Rafa Lopez

Partner: Robotnik Automation (Spain)

Inria contact: Iasonas Kokkinos

Type: FP7

Defi: Cognitive Systems and Robotics

Instrument: Specific Targeted Research Project

Objectif: Cognitive Systems and Robotics

Duration: February 2013 - January 2016

Coordinator: Dimos Dimarogonas

Partner: KTH (SE)

Inria contact: Iasonas Kokkinos

Type: FP7

Instrument: Career Integration Grant

Duration: January 2014 - December 2017

Coordinator: Inria

Inria contact: Matthew Blaschko

Title: Self-Paced Learning for Exploiting Noisy, Diverse or Incomplete Data

International Partner (Institution - Laboratory - Researcher):

Stanford University (ÉTATS-UNIS)

Duration: 2012 - 2014

See also: http://

The goal of the project is to develop methods for learning accurate probabilistic models using diverse (consisting of fully and weakly supervised samples), incomplete (consisting of partially labeled samples) and noisy (consisting of mislabeled samples) data. To this end, we will build on the intuitions gained from self-paced human learning, where a child is first taught simple concepts using simple examples, and gradually increasing the complexity of the concepts and the examples. In the context of machine learning, we aim to impart the learner with the ability to iteratively adapt the model complexity and process the training data in a meaningful order. The efficacy of the developed methods will be tested on several real world computer vision and medical imaging applications using large, inexpensively assembled datasets.

Europe

Technical University of Munich (DE) – Collaborative research with the Chair for Computer Aided Medical Procedures& Augmented Reality at the department of Computer Science. Collaboration Topic: Graph-based methods for linear/deformable registration, segmentation, and tracking.

University College London (UK) – Collaborative research with the Gatsby Computational Neuroscience Unit. Collaboration Topic: Kernel measures of dependence.

University of Oxford (UK) – Collaborative research with the Visual Geometry Group of the Department of Engineering Science. Collaboration Topic: Structured prediction and parts-based models.

University of Oulu (Finland) – Collaborative research with the Machine Vision Group at the department of Electrical Engineering. Collaboration Topic: Ranking based learning algorithms for cascaded object detection.

Americas

University of California at Los Angeles (US) – Collaborative research with the UCLA Vision Lab and the UCLA Center for Cognition, Vision, and Learning Lab at the Departments of Computer Science and Statistics. Collaboration Topic: Action Recognition & Object Detection Parsing.

University of Pensylvania (USA) – Collaborative research with the section of Biomedical Imaging of the Department of Radiology. Collaboration Topic: Graph-based methods for linear/deformable registration.

StonyBrook University, Computer Science Department (USA) – Collaborative research with the image analysis lab in the context of the SubSample DIGITEO Chair. Collaboration Topic: Higher Order Graph-based methods in graph-matching, cocaine addiction analysis with sparse graph models, object detection and implicit 3D pose estimation

Ecole Polytechnique de Montreal (CA) – Collaborative research with the Canada Research Chair in Medical Imaging and Assisted Interventions. Collaboration Topic: Higher Order Graph-based methods in Spine Imaging

University of Colorado, Department of Computer Science (USA) - Research with the Autonomous Robotics & Perception Group. Collaboration topic: Large scale video segmentation using efficient approximations to a graph Laplacian.

Asia

International Institute of Information Technology, Hyderabad (India) – Collaborative research with Center for Visual Information Technology. Collaboration Topic: Average precision with weak supervision and self-paced learning for deep convolutional neural networks.

Professor Maragos, Petros: Technical University of Athens, GR (October 2014)

Gastounioti, Aimilia: Technical University of Athens, GR (from February until June 2014)

Trulls, Eduard: Universitat Politècnica de Catalunya, ES (from June until October 2014)

Vedantam, Shanmukha Ramakrishna: Virginia Tech, USA ( from June 2014 until August 2014)]

Ferrante, Enzo: Stanford University, USA (from June to September 2014)

Boussaid, Haithem: University of Pennsylvania, USA (from June to September 2014)

Togkas, Stavros: Oxford University, UK (from August to November 2014)

Blaschko, Matthew: Co-Organizer of Learning and inference in discrete graphical models tutorial, in conjunction with IEEE Computer Vision and Pattern Recognition (CVPR).

Kokkinos, Iasonas: Co-Organizer of BASes for Images and Surfaces (BASIS) tutorial, in conjunction with IEEE Computer Vision and Pattern Recognition (CVPR).

Paragios, Nikos: (i) Co-Organizer of Bayesian and grAphical Models for Biomedical Imaging (BAMBI) workshop, in conjunction with the Medical Image Computing and Computer Assisted Intervention (MICCAI), (ii) Co-Organizer of the Learning and inference in discrete graphical models tutorial, in conjunction with IEEE Computer Vision and Pattern Recognition (CVPR).

Blaschko, Matthew: Neural Information Processing Systems (NIPS), British Machine Vision Conference (BMVC), Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP).

Kumar, Pawan: Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP).

Paragios, Nikos: IEEE Computer Vision and Pattern Recognition (CVPR), Medical Image Computing and Computer Assisted Intervention (MICCAI).

Argyriou, Andreas: Neural Information Processing Systems (NIPS).

Blaschko, Matthew: Artificial Intelligence and Statistics (AISTATS), Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR), IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

Kokkinos, Iasonas: European Conference on Computer Vision (ECCV), IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Neural Information Processing Systems (NIPS), Artificial Intelligence and Statistics (AISTATS), Asian Conference on Computer Vision (ACCV).

Kumar, Pawan: European Conference on Computer Vision (ECCV), Advances in Neural Information Processing Systems (NIPS).

Paragios, Nikos: European Conference on Computer Vision (ECCV).

Paragios, Nikos: Computer Vision and Image Understanding Journal (CVIU).

Kumar, Pawan: Computer Vision and Image Understanding (CVIU).

Kokkinos, Iasonas: Image and Vision Computing Journal (IVC), Guest Editor Special Issue on Generative Models in Computer Vision - Computer Vision and Image Understanding Journal (CVIU).

Paragios, Nikos: Medical Image Analysis Journal (MedIA), SIAM Journal on Imaging Sciences, Guest Editor Special Issue on Discrete Graphical Models in Biomedical Image Analysis - Medical Image Analysis Journal (MedIA).

Kokkinos, Iasonas: International Journal of Computer Vision (IJCV), IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), IEEE Transactions on Image Processing (T-IP), Image and Vision Computing (IVC), Computer Vision and Image Understanding (CVIU).

Kumar, Pawan: IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), Computer Vision and Image Understanding (CVIU).

Paragios, Nikos: International Journal of Computer Vision (IJCV), IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), IEEE Transactions on Medical Imaging (T-MI).

**Masters**

Blaschko, Matthew

Master: Foundations of Machine Learning, 36, M1, Ecole Centrale Paris, FR

Master: Structured Prediction, 24, M2, Ecole Centrale Paris, FR

Kokkinos, Iasonas

Master: Machine Learning for Computer Vision, 24, M2, Ecole Normale Superieure-Cachan, FR

Master: Introduction to Deep Learning, 24, M2, Ecole Centrale de Paris, FR

Master: Introduction to Signal Processing, 36, M1, Ecole Centrale de Paris, FR

Master: Introduction to Computer Vision, 36, M1, Ecole Centrale de Paris, FR

Kumar, Pawan

Master: Introduction to Discrete Optimization, 12, M2, Ecole Centrale de Paris, FR

Master: Discrete Optimization and Learning, 12, M2, Ecole Normale Superieure-Cachan, FR

Paragios, Nikos

Master: Advanced Mathematical Models in Computer Vision, 24, M2, Ecole Normale Superieure-Cachan, FR

**E-learning**

MOOC: Coursera

Pedagogical resources : Kumar, Pawan & Paragios, Nikos, Discrete Inference and Lerning in Artificial Vision, M2, https://

HdR : Matthew Blaschko, Advances in Empirical Risk Minimization for Image Analysis and Pattern Recognition, École Normale Supérieure de Cachan, 7 novembre 2014

PhD in progress : Puneet Kumar Dokania, Learning to Rank with Missing and High-Order Information, 2012-2015, M. Pawan Kumar

PhD in progress : Diane Bouchacourt, Large Scale Diverse Learning for Structured Output Prediction, 2014-2017, M. Pawan Kumar

PhD in progress: Haithem Boussaid, Efficient Inference and Learning in Graphical Models for Multi-organ Shape Segmentation, 2011-2015, I. Kokkinos

PhD in progress: Stavros Tsogkas, Learning structured mid-level representations for object recognition, 2011-2015, I. Kokkinos

PhD in progress: Siddhartha Chandra, Efficient Learning and Optimization for 3D Visual Data, 2013-2016, Iasonas Kokkinos, Pawan Kumar

PhD in progress: Stefan Kinauer, Surface-based representations for high-level vision tasks, 2013-2016, Iasonas Kokkinos.

PhD in progress : Wacha Bounliphone, Statistical tools for Imaging-Genetics data integration, 2013-2016, Matthew Blaschko & Arthur Tenenhaus

PhD in progress : Jiaqian Yu, Structured Prediction Methods for Computer Vision and Medical Imaging, 2014-2017, Matthew Blaschko

PhD in progress : Eugene Belilovsky, Structured Output Prediction on Large Scale Neuroscience Data, 2014-2017, Matthew Blaschko

PhD in progress : Stavros Alchatzidis, Message Passing Methods, Parallel Architectures & Visual Processing, 2011-2014, Nikos Paragios

PhD in progress : Enzo Ferrante, 2D-to-3D Multi-Modal Deformable Image Fusion, 2012-2015, Nikos Paragios

PhD in progress : Vivien Fecamp, Linear-Deformable Multi-Modal Deformable Image Fusion, 2012-2015, Nikos Paragios

PhD in progress : Evgenios Kornaropoulos, Diffusion Coefficient: a novel computer aided bio-marker, 2010-2013, Nikos Paragios

PhD in progress : Maxim Berman, Learning Higher Order Graphical Models, 2014-2017, Nikos Paragios

PhD in progress : Hariprasad Kannan, Efficient Inference on Higher Order Graphs, 2014-2017, Nikos Paragios

PhD in progress : Huu Dien Khue Le, Graph-based Visual Perception : Theories and Applications, 2014-2017, Nikos Paragios

**Matthew Blaschko**

**PhD Thesis Participation**: K. Gkirtzou - FR (PhD).

**Grant Reviewing Services**: European Research Council (ERC).

**Iasonas Kokkinos**

**PhD Thesis Participation**: N. Dimitriou - GR (PhD).

**Grant Reviewing Services:** Swiss National Science Foundation.

**Kumar, Pawan**

**PhD Thesis Participation**: K. Park - Australia (PhD), G. Lin - Australia (PhD).

**Paragios, Nikos**

**PhD Thesis Participation**: M. Blascho - FR (PhD), D. Fortun - FR (PhD), A. Gastounioti - GR (PhD), B. Romain - FR (PhD), J. Tang - CA (PhD), J. Weissenberg - CH (PhD).

**Grant Reviewing Services:** Agence National de la Recherche, Austrian Research Council, Danish Research Council, Dutch Research Council, European Research Council, Israel Research Foundation, Swiss National Science Foundation.

**Blaschko, Matthew**

**Presentations**: Third School on Machine Learning and Knowledge Discovery in Databases (BR),
Computer Vision and Pattern Recognition Tutorial (US), KU Leuven (BE), Machine Learning Challenge MICCAI Workshop (US),
Agence Nationale de la Recherche (FR)

**Kokkinos, Iasonas**

**Presentations**: Imagenet workshop (in conjunction with ECCV, CH), TTI-Chicago (USA), KTH University (SE), Dagstuhl Seminar on Shape Analysis (DE).

**Kumar, Pawan M.**

**Presentations**: University of Oxford (UK), Ecole des Ponts (FR), Swedish AI Society Workshop (SAIS '14, SE), Xerox Research Center Europe (XRCE) (FR).

**Paragios, Nikos**

**Presentations**: Reconnaissance de Formes et l'Intelligence Artificielle (RFIA'14, FR), Medical Imaging Summer School (MISS'14, IT), International Conference on Pattern Recogntion (ICPR'15, SE), Algorithmic issues for Inference in Graphical Models (AIGM'14, FR), University of Patras (GR), Swiss Federal Institute of Technology in Zurich (ETHZ) (CH).