News

The article ‘Crowdsourcing Biodiversity Monitoring: How Sharing your Photo Stream can Sustain our Planet‘ by Alexis Joly et al. has been added to the ACM Digital Library.

The article is part of the MM’16: Proceedings of the 2016 ACM on Multimedia Conference.

Abstract
Large scale biodiversity monitoring is essential for sustainable development (earth stewardship). With the recent advances in computer vision, we see the emergence of more and more effective identification tools allowing to set-up large-scale data collection platforms such as the popular Pl@ntNet initiative that allow to reuse interaction data. Although it covers only a fraction of the world flora, this platform is already being used by more than 300K people who produce tens of thousands of validated plant observations each year. This explicitly shared and validated data is only the tip of the iceberg. The real potential relies on the millions of raw image queries submitted by the users of the mobile application for which there is no human validation. People make such requests to get information on a plant along a hike or something they find in their garden but not know anything about. Allowing the exploitation of such contents in a fully automatic way could scale up the world-wide collection of implicit plant observations by several orders of magnitude, which can complement the explicit monitoring efforts. In this paper, we first survey existing automated plant identification systems through a five-year synthesis of the PlantCLEF benchmark and an impact study of the Pl@ntNet platform. We then focus on the implicit monitoring scenario and discuss related research challenges at the frontier of computer science and biodiversity studies. Finally, we discuss the results of a preliminary study focused on implicit monitoring of invasive species in mobile search logs. We show that the results are promising but that there is room for improvement before being able to automatically share implicit observations within international platforms.

The SLDESUTO team (ContextVision/HES-SO) obtained the 3rd place in two tasks of the ‘Tumor Proliferation Assessment Challenge (TUPAC 2016)’. The team used an automated deep learning approach to analyse whole slide digital pathology images. 
The corresponding tasks were:

• Prediction of the proliferation score based on mitosis counting

• Mitosis detection

More information can be found here

Sebastian Otalora has joined the Medgift group.

Welcome to him!

An scholarship ‘Bilateral research collaboration with the Asia-Pacific region 2013-2016’ funded by the Swiss State Secretariat for Education, Research and Innovation has been approved for our group. 

This scholarship from the Swiss confederation is targeted for projects that colaborate with partners in Malaysia. 

The current Microsoft Azure for research award has been increased. The VISCERAL benchmarks as well as other MedGIFT projects are currently sponsored by Azure Research.

For more information regarding the open VISCERAL benchmarks (Anatomy3 continuous, Detection2, Retrieval2) go to www.visceral.eu

The article “Deep learning with convolutional neural networks: a resource for the control of robotic prosthetic hands via electromyography”, by Manfredo Atzori, Matteo Cognolato, Henning Müller,  has now been published in Frontiers in Neurorobotics.

ABSTRACT

Natural control methods based on surface electromyography (sEMG) and pattern recognition are promising for hand prosthetics. However, the control robustness offered by scientific research is still not sufficient for many real life applications, and commercial prostheses are capable of offering natural control for only a few movements. In recent years deep learning revolutionized several fields of machine learning, including computer vision and speech recognition. Our objective is to test its methods for natural control of robotic hands via sEMG using a large number of intact subjects and amputees. We tested convolutional networks for the classification of an average of 50 hand movements in 67 intact subjects and 11 transradial amputees. The simple architecture of the neural network allowed to make several tests in order to evaluate the effect of pre-processing, layer architecture, data augmentation and optimization. The classification results are compared with a set of classical classification methods applied on the same datasets. The classification accuracy obtained with convolutional neural networks using the proposed architecture is higher than the average results obtained with the classical classification methods, but lower than the results obtained with the best reference methods in our tests. The results show that convolutional neural networks with a very simple architecture can produce accurate results comparable to the average classical classification methods. They show that several factors (including pre-processing, the architecture of the net and the optimization parameters) can be fundamental for the analysis of sEMG data. Larger networks can achieve higher accuracy on computer vision and object recognition tasks. This fact suggests that it may be interesting to evaluate if larger networks can increase sEMG classification accuracy too.

Prof. Henning Müller will be an invited speaker at the workshop:  Challenges in Machine Learning: Gaming and Education (CiML) 2016. This is one of the NIPS workshops and will take place Friday December 9, 2016, Barcelona, Spain. 

Motivations

Challenges in machine learning and data science are competitions running over several weeks or months to resolve problems using provided datasets or simulated environments. 

The playful nature of challenges naturally attracts students, making challenge a great teaching resource. For this third edition of the CiML workshop at NIPS we want to explore more in depth the opportunities that challenges offer as teaching tools. The workshop will give a large part to discussions around several axes: (1) benefits and limitations of challenges to give students problem-solving skills and teach them best practices in machine learning; (2) challenges and continuous education and up-skilling in the enterprise; (3) design issues to make challenges more effective teaching aids; (3) curricula involving students in challenge design as a means of educating them about rigorous experimental design, reproducible research, and project leadership.
CiML is a forum that brings together workshop organizers, platform providers, and participants to discuss best practices in challenge organization and new methods and application opportunities to design high impact challenges. Following the success of last year’s workshop, in which a fruitful exchange led to many innovations, we propose to reconvene and discuss new opportunities for challenges in education, one of the hottest topics identified in last year’s discussions. We have invited prominent speakers in this field. 
We will also reserve time to an open discussion to dig into other topic including open innovation, coopetitions, platform interoperability, and tool mutualisation.

Prof. Adrien Depeursinge was a special guest at the QTIM meeting in Boston last July 27, 2016. The topic of his talk was ‘A 3-D Riesz-Covariance Texture Model for the Prediction of Nodule Recurrence in Lung CT’ 

Abstract:
We propose a novel imaging biomarker of lung cancer relapse from 3-D texture analysis of CT images. Three-dimensional morphological nodular tissue properties are described in terms of 3-D Riesz-wavelets. The responses of the latter are aggregated within nodular regions by means of feature covariances, which leverage rich intra- and inter-variations of the feature space dimensions. The obtained Riesz-covariance descriptors lie on a manifold governed by Riemannian geometry requiring specific geodesic metrics to locally approximate scalar products. The latter are used to construct a kernel for support vector machines (SVM). The effectiveness of the presented models is evaluated on a dataset of 92 patients with non-small cell lung carcinoma (NSCLC) and cancer recurrence information. Disease recurrence within a timeframe of 12 months could be predicted with an accuracy above 80, and highlighted the importance of covariance-based texture aggregation. At the end of the talk, computer tools will be presented to easily extract 3D radiomics quantitative features from PET-CT images.

The CLEF2016 Working Notes are now on-line on CEUR-WS Proceedings: http://ceur-ws.org/Vol-1609/

The Working Notes correspond to the Conference and Labs of the Evaluation forum that took place in Évora, Portugal, 5-8 September, 2016.

Adrien Depeursinge will be a speaker at the MIT CSAIL Seminar Series: Biomedical Imaging and Analysis 2015/2016. The topic of his talk is ‘Steerable Wavelet Machines (SWM): Learning Moving Frames for Texture Classification’. The presentation will take place Thursday, July 28, 2016.

SUMMARY
We present texture operators encoding class-specific local organizations of image directions (LOID) in a rotation-invariant fashion. The LOIDs are key for visual understanding, and are at the origin of the success of the popular approaches such as local binary patterns (LBP) and the scale-invariant feature transform (SIFT). Whereas LBPs and SIFT yield handcrafted image representations, we propose to learn data-specific representations of the LOIDs in a rotation-invariant fashion. The image operators are based on steerable circular harmonic wavelets (CHW), offering a rich and yet compact initial representation for characterizing natural textures. The joint location and orientation required to encode the LOIDs is preserved by using moving frames (MF) texture representations built from locally-steered multi-order CHWs. In a second step, we use support vector machines (SVM) to learn a multi-class shaping matrix of the initial CHW representation, yielding data-driven MFs that are invariant to rigid motions. We experimentally demonstrate the effectiveness of the proposed operators for classifying natural textures.

Henning Müller was teaching in a workshop of the Harvard/MIT Health Science and Technology program at MIT on July 13. The workshop was on Topics in clinical imaging informatics: beyond the radiology report.

The Ninapro project has the following updates:

1) A Ninapro summary paper was accepted on the Italian Journal of Hand Surgery and it will be published in September. If needed, you can download the paper from the following link: https://www.dropbox.com/s/w8fxhbthzmljhg6/Atzori_RCM_2016.doc?dl=0

2) 2) An electronic poster was presented the XXI Congress of the Federation of European Societies for Surgery of the Hand, in Santander: http://publications.hevs.ch/index.php/publications/show/2095

3) The paper ‘Clinical Parameter Effect on the Capability to Control Myoelectric Robotic Prosthetic Hands’ by Manfredo Atzori et al. was published in May in the Journal of Rehabilitation Research and Development: http://publications.hevs.ch/index.php/publications/show/1981

The paper ‘Cloud-based Evaluation of Anatomical Structure Segmentation and Landmark Detection Algorithms: VISCERAL Anatomy Benchmarks’ by Oscar Jimenez-del-Toro et al. is now available online in IEEE Xplore.

More information about the VISCERAL project can be found at www.visceral.eu

ABSTRACT

Variations in the shape and appearance of anatomical structures in medical images are often relevant radiological signs of disease. Automatic tools can help automate parts of this manual process. A cloud-based evaluation framework is presented in this paper including results of benchmarking current state-of-the-art medical imaging algorithms for anatomical structure segmentation and landmark detection: the VISCERAL Anatomy benchmarks. The algorithms are implemented in virtual machines in the cloud where participants can only access the training data and can be run privately by the benchmark administrators to objectively compare their performance in an unseen common test set. Overall, 120 computed tomography and magnetic resonance patient volumes were manually annotated to create a standard Gold Corpus containing a total of 1295 structures and 1760 landmarks. Ten participants contributed with automatic algorithms for the organ segmentation task, and three for the landmark localization task. Different algorithms obtained the best scores in the four available imaging modalities and for subsets of anatomical structures. The annotation framework, resulting data set, evaluation setup, results and performance analysis from the three VISCERAL Anatomy benchmarks are presented in this article. Both the VISCERAL data set and Silver Corpus generated with the fusion of the participant algorithms on a larger set of non-manually-annotated medical images are available to the research community.

The 14th International Workshop on Content-based Multimedia Indexing , June 15-17 2016, took place in Bucharest, Romania. The event was a success and brough together various communities involved in all aspects of content-based multimedia indexing for retrieval, browsing, visualization and analytics. 

The VISCERAL project received a Microsoft Azure sponsorship until June 2017. This will allow us to continue running the VISCERAL benchmarks in the Azure cloud platform for an extended period of time.  

The current open benchmarks are:

Anatomy3 continuous: The Benchmark tasks are the segmentation of bones, inner organs and relevant substructures. The Benchmark runs under a continuous evaluation approach – results on the test set can be obtained at any time.

Detection2: The goal is to automatically detect lesions in images. We will distribute training data with a large number of expert annotated lesions for training of detection algorithms. During the evaluation these algorithms will be applied to new data. The lesions are in bones, lungs, brain, liver, lymphnodes.

Retrieval2: The objective of retrieval algorithms will be to find cases with similar anomalies based on query cases.

The article “Optimized Distributed Hyperparameter Search and Simulation for Lung Texture Classification in CT Using Hadoop” by Roger Schaer et al. is now available online: http://www.mdpi.com/2313-433X/2/2/19/html

Abstract
Many medical image analysis tasks require complex learning strategies to reach a quality of image-based decision support that is sufficient in clinical practice. The analysis of medical texture in tomographic images, for example of lung tissue, is no exception. Via a learning framework, very good classification accuracy can be obtained, but several parameters need to be optimized. This article describes a practical framework for efficient distributed parameter optimization. The proposed solutions are applicable for many research groups with heterogeneous computing infrastructures and for various machine learning algorithms. These infrastructures can easily be connected via distributed computation frameworks. We use the Hadoop framework to run and distribute both grid and random search strategies for hyperparameter optimization and cross-validations on a cluster of 21 nodes composed of desktop computers and servers. We show that significant speedups of up to 364× compared to a serial execution can be achieved using our in-house Hadoop cluster by distributing the computation and automatically pruning the search space while still identifying the best-performing parameter combinations. To the best of our knowledge, this is the first article presenting practical results in detail for complex data analysis tasks on such a heterogeneous infrastructure together with a linked simulation framework that allows for computing resource planning. The results are directly applicable in many scenarios and allow implementing an efficient and effective strategy for medical (image) data analysis and related learning approaches.

The ‘Report on the Cloud-based Evaluation Approaches Workshop 2015’ by Henning Müller et al. is now available online: http://sigir.org/files/forum/2016J/p035.pdf

Prof. Henning Müller is Series Editor of the new Springer series Brain Informatics and Health. The website of the book is now online.

The second edition of the ‘ArkAthon Hacking Health Valais‘ will take place 3-5 june 2016 at Sierre, Switzerland

This event will start at the end of the eHealth Day (Friday, June 3). Participating teams will have 48 hours to solve the challenges on the platform trying to reach concrete results and proposals. At its conclusion, a jury will decide on the projects presented in the competition. The winning project will win the main prize with a total value of CHF 25,000.

The paper ‘Medical image modality classification using discrete Bayesian networks’ by Jacinto Arias, Jesus Martínez-Gómez, Jose A. Gámez, Alba G. Seco de Herrera and Henning Müller is now publised online.

ABSTRACT

In this paper we propose a complete pipeline for medical image modality classification focused on the application of discrete Bayesian network classifiers. Modality refers to the categorization of biomedical images from the literature according to a previously defined set of image types, such as X-ray, graph or gene sequence. We describe an extensive pipeline starting with feature extraction from images, data combination, pre-processing and a range of different classification techniques and models. We study the expressive power of several image descriptors along with supervised discretization and feature selection to show the performance of discrete Bayesian networks compared to the usual deterministic classifiers used in image classification. We perform an exhaustive experimentation by using the ImageCLEFmed 2013 collection. This problem presents a high number of classes so we propose several hierarchical approaches. In a first set of experiments we evaluate a wide range of parameters for our pipeline along with several classification models. Finally, we perform a comparison by setting up the competition environment between our selected approaches and the best ones of the original competition. Results show that the Bayesian Network classifiers obtain very competitive results. Furthermore, the proposed approach is stable and it can be applied to other problems that present inherent hierarchical structures of classes.