Spatial statistics and point processes, have been a fundamental field of research for decades. Leading methods in image processing rely on powerful image models/priors, e.g., marked point processes and random sets. They are now especially appropriate to analyze the spatial distribution of proteins or single molecules observed in fluorescence microscopy and super-resolution imaging. Recently, they have been utilized for solving inverse problems in bioimaging (e.g. deconvolution), tracking moving particles or co-localisation in fluorescence microsopy. Machine learning techniques and convolutional neural networks are now investigated to address similar issues. The proposed minisymposium consists of two sessions, covering a series of problems in this field.

GcoPS: a fast automatic colocalization method for 3D live cell imaging and super-resolution microscopy

Charles Kervrann (Inria)

Marked point processes for detecting objects in microscopy

Descombes Xavier (Inria Sophia-Antipolis)

Spatial statistics extends co-localization analysis to non-local interaction analysis

Ivo Sbalzarini (TU Dresden / Max Planck Insitute of Molecular Cell Biology and Genetics)

Spatial patterns in large-scale bioimaging data: applications in spatial transcriptomics and phenomics

Walter Thomas (CBIO, Mines ParisTech)

Organizers:

Charles Kervrann (Inria)

Keywords:

bayesian methods, deep learning, detection, image registration, image representation, image segmentation, inverse problems, machine learning, segmentation, spatial statistics, stochastic geometry, stochastic processes