Automated pattern recognition for live cell microscopy: the cost of blind faith and how to avoid it

Philippe Roudot
I2M, Aix-Marseille Université
/user/philippe.roudot/

Date(s) : 15/03/2021   iCal
13 h 30 min - 14 h 30 min

Continuous improvements in three-dimensional fluorescence microscopy have provided a new window into the heterogeneity of biological processes in systems of growing physiological relevance. To match the growing complexity of such datasets, new pattern recognition approaches must be investigated to ensure the automated and error-controlled detection of patterns that may not be observable a priori.  In this talk, we will focus on the problem of measurement error control in the context of multiple particle tracking for 3D light-sheet microscopy. We will describe how local error estimates and error propagation techniques can provide rich insights on measurement quality throughout the volumetric sequence. We will also show how these technologies can be exploited to discover measurable processes of interest, to drive the design of quantitative experiments and to help with the calibration of microscope.
 
This seminar will begin by a short internal seminar, given by Shao-Zen Lin, CENTURI postdoc (CPT), entitled “Theoretical modeling of active tissue flows”.  Short internal seminars are a great opportunity to discover the work of CENTURI PhD students and postdoctoral fellows.
 
This seminar will be followed by a tutored seminar with Philippe Roudot. We have changed the format of these tutored seminars: they are now an informal discussion with Philippe Roudot, about the content of the seminar, scientific careers, etc. If you wish to participate in this tutored seminar, please contact Mélina De Oliveira at the following address: melina.de-oliveira@univ-amu.fr.
 
Please note that the seminar will take place online via Zoom.
 
WEBINAR ID NUMBER: 864 4887 3722
PASSWORD: see mail

Philippe Roudot: about his research

Catégories



Retour en haut 

Secured By miniOrange