Our Scientific Data article on “A large calibrated database of hand movements and grasps kinematics“, by Néstor J. Jarque-Bou, Manfredo Atzori and Henning Müller is now available here.
Vincent Andrearczyk presented our work on exploiting images from public biomedical literature for machine learning at the 26th international conference on MultiMedia Modelling (MMM2020) in Daejeon, Korea.
Full details of the paper: “Studying Public Medical Images from the Open Access Literature and Social Networks for Model Training and Knowledge Extraction” by Henning Müller, Vincent Andrearczyk, Oscar Jimenez del Toro, Anjani Dhrangadhariya, Roger Schaer and Manfredo Atzori.
The paper entitled “A large calibrated database of hand movements and grasps kinematics“, by Néstor Jarque-Bou, Manfredo Atzori, and Henning Müller, has been accepted for publication in Scientific Data, a journal from the publishers of Nature. The manuscript will soon be available online.
Prof. Henning Müller gave invited talks at the Conference on Artificial Intelligence in Medical Imaging, Tehran, Nov. 2-5 2019.
He gave a presentation on “Creating an AI ecosystem for multimodal data analysis in the MedGIFT group” and an overview of the VISCERAL benchmarks on intelligent medical data analysis.
Valentin Oreiller gave an invited talk on “Learning with less data: Parametric and Locally Rotation Invariant CNNs” at the Workshop on Machine Learning in Radiology 2019 in CHUV (Centre Hospitalier Universaire Vaudois), Lausanne.
This workshop, co-organized by CHUV Department of Radiology and Siemens Healthineers Advanced Clinical Imaging Technology, gathered national and international researchers for discussions and talks on the use of machine learning with radiology data, and how these techniques can benefit clinical practice.
Prof. Henning Müller is joining the editorial board of Scientific Data, a data-focused journal from the publishers of Nature.
Scientific Data is a peer-reviewed open-access journal focusing on descriptions of data sets relevant to the natural sciences, which are provided as machine-readable data, complemented with a human-oriented narrative.
Four papers have been accepted for oral presentations at the 2020 SPIE Medical Imaging conference which will be held in Houston, Texas, 15 – 20 February 2020.
“Exploiting the PubMed central repository to mine out a large multimodal dataset of rare cancer studies” by Anjani K. Dhrangadhariya, Oscar Jimenez-del-Toro, Vincent Andrearczyk, Manfredo Atzori and Henning Müller, in the Imaging Informatics for Healthcare, Research, and Applications track.
“A systematic comparison of deep learning strategies for weakly supervised Gleason grading” by Sebastian Otálora, Oscar Jimenez-del-Toro, Manfredo Atzori, Amjad Khan, Vincent Andrearczyk and Henning Müller, in the Digital Pathology track.
“An exploration of uncertainty information for segmentation quality assessment” by Katharina Hoebel, Vincent Andrearczyk, Andrew L. Beers, Jay B. Patel, Ken Chang, Adrien Depeursinge, Henning Müller and Jayashree Kalpathy-Cramer, in the Image Processing track.
“Generalizing convolution neural networks on stain color heterogeneous data for computational pathology”, by Amjad Khan, Manfredo Atzori, Sebastian Otálora, Vincent Andrearczyk and Henning Müller, in the Digital Pathology track.
The paper entitled “Texture-driven parametric snakes for semi-automatic image segmentation” by Anaïs Badoual, Michael Unser and Adrien Depeursinge has been accepted for publication in Computer Vision and Image Understanding (CVIU).
Nora’s presentation Educational Outcomes Depend Both on Visual and Multisensory Control of Selective Attention has been selected as a flash talk for the 1st Annual Meeting of the Swiss Society of the Early Childhood Research, taking place on 18th of November 2019 in Lausanne.
23rd September: Paul gave an invited talk “Bridging past and future in education” at the Development-Neuroscience-Cognition section of Psychology Dept of UZH by invitation of Pr. Nora Raschle: https://www.psychologie.uzh.ch/de/studium/bscmsc/master/kolloquien/denc.html
29th Aug – 1st Sep: Nora Turoman (now employed at HES SO from 1st of October) & Paul took part in the annual meeting of FLUX developmental neuroscience Society. As part of this conference:
– Paul acted as a discussion facilitator of a workgroup Doing Research in Naturalistic Settings during the Flux 2019 Pre-Conference workshop “Beyond the lab: Translating developmental neuroscience” organised by Dr Lucia Magis Weinberg & Natasha Duell: https://twitter.com/paweljmatusz/status/1167068748618043393?s=20
– During the FLUX 2019 Conference, Paul also chaired a competitive poster symposium entitled “Lost in Translation: How does neuroimaging contribute to understanding development in everyday diverse environments”: https://twitter.com/paweljmatusz/status/1167862242190061569?s=20
– At the Flux 2019 Conference, Nora presented a poster on her latest findings on the dependence of educational outcomes on both visual and multisensory attentional control, as a function of age: https://twitter.com/NoraPlethora/status/1167456509049147393?s=20
The paper “Fast signal feature extraction using parallel time windows” by Manfredo Atzori and Henning Müller has been accepted for publication in Frontiers in Neurorobotics.
The paper entitled “Interpreting Intentionally Flawed Models with Linear Probes” by Mara Graziani, Henning Müller and Vincent Andrearczyk has been accepted for presentation at The First Workshop on Statistical Deep Learning in Computer Vision (SDL-CV). The workshop will be held in conjunction with ICCV 2019 in Seoul, Korea 27th Oct.
Find out more about the paper and source code on the medium post at this link.
Our new paper on staining invariant features in deep CNNs has been accepted for publication in Frontiers in Bioengineering and Biotechnology.
Paper now available at this link.
Title and authors: “Staining invariant features for improving generalization of deep convolutional neural networks in computational pathology”, Sebastian Otálora, Manfredo Atzori, Vincent Andrearczyk, Amjad Khan and Henning Müller
The paper Visualizing and Interpreting Feature Reuse of Pretrained CNNs for Histopathology by Mara Graziani, Vincent Andrearczyk and Henning Müller has been accepted at the Irish Machine Vision and Image Processing conference (IMVIP 2019). Meet our team at the poster presentation in Dublin, 28-30 August.
More details, paper preview and github repo in the medium post.
The paper “Solid Spherical Energy (SSE) CNNs for Efficient 3D Medical Image Analysis”, by Vincent Andrearczyk, Valentin Oreiller, Julien Fageot, Xavier Montet and Adrien Depeursinge has been accepted for presentation at the Irish Machine Vision and Image Processing conference (IMVIP 2019), to be held in Dublin 28-30 August.
Our paper on “Exploring Local Rotation Invariance in 3D CNNs With Steerable Filters” by Vincent Andrearczyk, Julien Fageot, Valentin Oreiller, Xavier Montet, Adrien Depeursinge obtained the best paper award at the International Conference on Medical Imaging with Deep Learning (MIDL 2019).
The oral presentation is available here.