Cell motility tracking of intravital microscopy via binary fitting energy driven model for level-set

摘要

Quantifying the motion and deformation of cells through image sequences obtain with microscopy is a recurrent task.Firstly,separate the clustering cells contour of blured image,a novel cell boundary feature extraction algorithm based on it uses a new kind of region representative: binary fitting energy is proposed.Binary fitting energy driven model allows for an excellent approximation of smooth region at global scope.Then contour strategy for digital cell image by local binary fitting energy under variational model framwork is presented.It is also combined a robust convergence criteria and a scheme to determine the optimal time-step for the numerical solution of the level set equation in this approach.It is more accurate than the classical contour extraction algorithm under variational model framework,for bio-inspired cell contour extraction corrupted by interference,with blured edges.Experimental results tested by different low visual quality of tissue cells demonstrate good performances of the proposed method for tracking quantification of cell motility.The model stays simple and relatively fast to compute.