Nonlinear PDE based numerical methods for cell tracking in zebrafish embryogenesis

The paper presents numerical algorithms leading to an automated cell tracking and reconstruction of the cell lineage tree during the first hours of animal embryogenesis. We present results obtained for large-scale 3D+time two-photon laser scanning microscopy images of early stages of zebrafish (Dani...

Full description

Saved in:
Bibliographic Details
Published in:Applied numerical mathematics 2015-09, Vol.95, p.250-266
Main Authors: Mikula, Karol, Špir, Róbert, Smíšek, Michal, Faure, Emmanuel, Peyriéras, Nadine
Format: Article
Language:English
Subjects:
Cell tracking
Cognitive Sciences
Image processing
Life Sciences
Neurobiology
Neurons and Cognition
Nonlinear partial differential equations
Numerical methods
Psychology and behavior
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The paper presents numerical algorithms leading to an automated cell tracking and reconstruction of the cell lineage tree during the first hours of animal embryogenesis. We present results obtained for large-scale 3D+time two-photon laser scanning microscopy images of early stages of zebrafish (Danio rerio) embryo development. Our approach consists of three basic steps – the image filtering, the cell centers detection and the cell trajectories extraction yielding the lineage tree reconstruction. In all three steps we use nonlinear partial differential equations. For the filtering the geodesic mean curvature flow in level set formulation is used, for the cell center detection the motion of level sets by a constant speed regularized by mean curvature flow is used and the solution of the eikonal equation is essential for the cell trajectories extraction. The core of our new tracking method is an original approach to cell trajectories extraction based on finding a continuous centered paths inside the spatio-temporal tree structures representing cell movement and divisions. Such paths are found by using a suitably designed distance function from cell centers detected in all time steps of the 3D+time image sequence and by a backtracking in the steepest descent direction of a potential field based on this distance function. We also present efficient and naturally parallelizable discretizations of the aforementioned nonlinear PDEs and discuss properties and results of our new tracking method on artificial and real 4D data.
ISSN:0168-9274
1873-5460
DOI:10.1016/j.apnum.2014.09.002