Deadline for full application: December 15th, 2013
Interviews: March, 2014
Start of the Ph.D.: October 1st, 2014
Department: Cell Biology and Infection
Title of the PhD project: Computerized Analysis of Animal Behavior
Name of the lab: Quantitative Image Analysis
Head of the lab: Jean-Christophe Olivo-Marin
PhD advisor: Jean-Christophe Olivo-Marin
Email address: firstname.lastname@example.org
Web site address of the lab: www.bioimageanalysis.org
Doctoral school affiliation and University: EDITE de Paris, Paris 6
Presentation of the laboratory and its research topics:
The Quantitative Image Analysis (QuIA) unit is dedicated to design innovative and rigorous quantitative image analysis and cosmputer vision methods for the processing and quantification of multi-channel temporal 3D sequences in biological imaging, especially at cellular and molecular levels.
Description of the project:
The overall rationale of this project is to decipher key spatiotemporal steps of mice behaviour in relation with cellular and molecular events through the systematic use of 2D and 3D+t imaging and computer vision. Deciphering the complex patterns of mice social behaviours necessitates indeed large-scale automated image-based protocols to document the wide range of highly variable and subtle characteristics of animal behaviour. We have shown (de Chaumont et al, 2012a) that extensive imaging of animals combined with sophisticated and powerful computational analysis of their motility and interaction patterns provide new insights contributing to a detailed description, quantification and understanding of behavioural schemes.
We now want to go further on and take advantage of recent advances in the field of computer vision and signal processing (de Chaumont et al, 2012b) to automate the analysis of recorded movies documenting the behaviour of mice in different setups. The major methodological goals of this project will be to: 1) extract automatically (meta)data describing the spatio-temporal trajectories of animals and the timing of behavioural events; and 2) combine specific video and audio (meta)data, extract analytical information and store it in databases for population analysis. A key aspect of this project will therefore be to develop various data analysis and image processing algorithms necessary for the tracking of mice and the subsequent correlation of video and audio data. The mathematical developments will address articulated body models for precise modelling and representation of animal movements, scene reconstruction from data fusion, and signal processing in cluttered environments for multi-animal detection and identification. Many of these developments are challenging frontier research topics in image/signal processing and computer vision that have not been fully developed in the context of animal behavioural studies.
By providing innovative tools that are mostly lacking today, this project will enable the automated processing and visualisation of temporal 2D/3D +t sequences and the integration of multimodality video and audio data with genetic information and functional assays. In collaboration with neuroscientists, this approach is expected to provide innovative and powerful tools to behavioural analysis. By enabling the efficient analysis of animal models for precise phenotypic traits, it will indeed allow combining insights about brain circuits with information about genes that may control one particular building block of specific complex behaviours.
Computer analysis, 3D visualization, tracking, multimodal analysis, animal behavior
Expected profile of the candidate (optional):
Computer Science, Physics, Applied Mathematics, Electrical Engineering
Jean-Christophe Olivo-Marin (email@example.com)