Daniel Sage was born in Annecy, France. He received the Master degree and Ph.D. degrees in signal and image processing from the Institut National Polytechnique de Grenoble INPG, France. He did his research Ph.D. thesis at the GIPSA laboratory (previously TIRF) on tracking methods. From 1989 to 1998, he was a Consulting Engineer developing vision systems for quality control, then Head of the Industrial Vision Department of Attexor S.A. During his career, he has developed some vision systems oriented to the quality control in the industrial sector.
In 1998, Daniel Sage joined the Biomedical Imaging Group (BIG) of the Prof. M. Unser at Ecole Polytechnique Fédérale de Lausanne (EPFL) as responsible of the Head of the Software Development. He is currently in charge of the support to the researchers of the laboratory and also to the research community of the EPFL Center for Imaging. He is involved in numerous research projects in computational bioimaging including super-resolution microscopy, tracking, deconvolution, and image quantification. He is engaged in the open-source software development for the life science community, using both engineering and machine learning methods. He is also involved in the teaching of image processing and image analysis, including the development of methods for computer-assisted teaching.
User friendly tools for deep-learning bioimage analysis: Hands-on DeepImageJ (Part I)
Summary of the course
Machine learning (ML), and in particular, Deep Neural Networks (DNN), have become an inflection point in many areas of scientific research. In the case of biomedical image analysis, these new techniques provide significant improvements in most of the tasks such as denoising, super-resolution, segmentation, detection, tracking, response prediction or computer-aided diagnosis.
Nonetheless, the use of DNN models requires previous programming knowledge, making its use a challenging task for non-expert users. This problem limits the impact of these revolutionary methods. However, some softwares are starting to build a bridge between the developers of DNN and their target users. In order to contribute to this effort DeepImageJ appeared. In this practical session we are going to talk about DeepImageJ, a new plugin to run pre- trained Deep Learning models in ImageJ.
The plugin is developed in a user-friendly manner that allows running models with few clicks and no code, thus combining DNN with other powerful Bioimage Analysis routines already available in ImageJ. DeepImageJ also integrates the BioImage.io model zoo directly into its interface, making a growing collection of Deep Learning models readily available.
During the practical session we will explore all the mentioned features of the plugin in addition to other novel ones, such as using a server as the backend of the model inference, which improves the speed and performance of the plugin greatly.
Main topic: Introduction on DL + Bioimage Model Zoo + Prediction deepImageJ + U-net training on ZeroCost (HeLa cell segmentation)
- deepImageJ: https://www.nature.com/articles/s41592-021-01262-9
- ZeroCostDL4Mic: https://www.nature.com/articles/s41467-021-22518-0
- Bioimage Model Zoo: https://www.biorxiv.org/content/10.1101/2022.06.07.495102v1