Presentations at the Valais/Wallis AI Workshop on interpretability

The 5th Valais/Wallis AI Workshop took place last Friday May 3rd at the Idiap Research Institute.

Mara Graziani presented her work on “Concept Measures to Explain Deep Learning Predictions in Medical Imaging” (video).

Vincent Andrearczyk talked about the “Transparency of rotation-equivariant CNNs via local geometric priors” (video).

Keynote speech by Prof. Pena Carlos Andrés and other presentations are also available here.

Paper accepted at TAROS 2019

The paper entitled “An Augmented Reality Environment to Provide Visual Feedback to Amputees during sEMG Data Acquisitions” by Francesca Palermo, Matteo Cognolato, Ivan Eggel, Manfredo Atzori and Henning Müller has been accepted for publication at Towards Autonomous Robotic Systems (TAROS 2019). The 20th TAROS conference will be held at the Queen Mary University of London, 3-5 July 2019.

Paper on lung graph models accepted at IPMI 2019

The paper entitled “A Graph Model of the Lungs with Morphology-based Structure for Tuberculosis Type Classification” by Yashin Dicente Cid, Oscar Jiménez-del-Toro, Pierre-Alexandre Poletti and Henning Müller, has been accepted at the 26th biennial international conference on Information Processing in Medical Imaging (IPMI 2019). The conference will be held in Hong Kong in June (2-7).

Congratulations to the authors!

Paper accepted in Nature Communications

The following paper  has been accepted for publication in Nature Communications:

Why rankings of biomedical image analysis competitions should be interpreted with care“, by Lena Maier-Hein, Matthias Eisenmann, Annika Reinke, Sinan Onogur, Marko Stankovic, Patrick Scholz, Tal Arbel, Hrvoje Bogunovic, Andrew Bradley, Aaron Carass, Carolin Feldmann, Alejandro Frangi, Peter Full, Bram van Ginneken, Allan Hanbury, Katrin Honauer, Michal Kozubek, Bennett Landman, Keno März, Oskar Maier, Klaus Maier-Hein, Bjoern Menze, Henning Müller, Peter Neher, Wiro Niessen, Nasir Rajpoot, Gregory Sharp, Korsuk Sirinukunwattana, Stefanie Speidel, Christian Stock, Danail Stoyanov, Abdel Aziz Taha, Fons van der Sommen, Ching-Wei Wang, Marc-André Weber, Guoyan Zheng, Pierre Jannin, and Annette Kopp-Schneider

Three papers accepted at SPIE Medical Imaging 2019

The following three papers have been accepted at SPIE Medical Imaging, to be held in San Diego, California, United States, 16 – 21 February 2019.
  • “Improved interpretability for computer-aided severity assessment of retinopathy of prematurity “, Mara Graziani, James M. Brown, Vincent Andrearczyk, Veysi Yildiz, J Peter Campbell, Deniz Erdogmus, Stratis Ioannidis, Michael F. Chiang, Jayashree Kalpathy-Cramer and Henning Müller, oral presentation at the Computer-Aided Diagnosis conference track.
  • “Learning cross-protocol radiomics and deep feature standardization from CT images of texture phantoms”, Vincent Andrearczyk, Adrien Depeursinge and Henning Müller,  oral presentation at the Imaging Informatics for Healthcare, Research, and Applications conference track.
  • “A Lung Graph Model for the Classification of Interstitial Lung Disease on CT Images”, Guillaume Vanoost, Yashin Dicente and Adrien Depeursinge, poster presentation at the Computer-Aided Diagnosis conference track.

Best paper award at MICCAI iMIMIC

Congratulations to Mara Graziani for receiving the best paper award for her work on “Regression Concept Vectors for Bidirectional Explanations in Histopathology” at MICCAI iMIMIC workshop, Granada, 2018.

Regression Concept Vectors help in evaluating the importance of concept measures such as morphological or texture features in the decisions of deep neural networks. Bidirectional relevance scores can be used to get insights and unpack the deep learning black box.

paper preprint

 

Project accepted in collaboration with the institute for information problems at the National Academy of Sciences, Belarus

The project “Exploring the utility of new Artificial Intelligence and Deep Learning methods for radiological image analysis and lung disease diagnosis” was accepted by the  Leading House for Swiss-Russian Science & Technology Cooperation. Principal applicant prof. Henning Müller, in collaboration with the institute for information problems at the National Academy of Sciences, Belarus

EXA MODE Project passed the ICT-12-2018-2020 evaluation phase

The EXA MODE Project passed the ICT-12-2018-2020 — Big Data technologies and extreme-scale analytics call evaluation phase.
EXA MODE (EXtreme-scale Analytics via Multimodal Ontology Discovery & Enhancement) includes seven partners from the European Union and Switzerland. The project will be financed by EU in order to revolutionise digital pathology.

Invited talk at the MICCAI COMPAY workshop

On Sepember 16, at the Granada Conference Center, Dr. Manfredo Atzori will present the talk “Data variability as a challenge to improve classification and retrieval in digital pathology” at the workshop on Computational Pathology (COMPAY) part of the 21st international conference on medical image computing & computer assisted intervention (MICCAI 2018).

link to invited speakers: https://sites.google.com/site/compaysymposium2018/speakers