Title : “Statistical and Deep Learning for Image Reconstruction in Microscopy: Performance, Limitations, and Perspectives”
Charles Kervrann, from Inria Centre at the University of Rennes (SAIRPICO Project Team & Institut Curie) will give a seminar on statistical and deep learning methods for image reconstruction in fluorescence microscopy on Wednesday, July 15, at 04:00 p.m. in Room 010 (meeting room ground floor) at Institut Fresnel, Marseille.
Abstract : Deep learning has profoundly impacted image reconstruction in fluorescence microscopy, enabling major advances in denoising, deconvolution, and image reconstruction. In this seminar, Charles Kervrann will present the foundations of convolutional neural networks and statistical learning approaches for these applications. Beyond their impressive performance, he will discuss important practical questions regarding interpretability, robustness, and generalization, and compare these methods with more traditional image processing techniques that remain highly competitive in many imaging scenarios.
