Quantitative characterization for lymphocyte deformation in image sequences

A new method is proposed to quantitatively characterize lymphocyte deformation in image sequences. The whole framework includes three steps:cell segmentation, cell tracking, and cell deformation characterization. Firstly, morphological image processing is implemented as a kind of segmentation method to detect cell region. Then, method based on Fourier descriptors is presented to optimize cell contour, which makes cell contour smooth and closer to real contour. Secondly, gradient vector flow (GVF) Snake algorithm is carried out for tracking cell region in image sequences. Finally, the part of cell deformation characterization is accomplished by improved symmetry difference for cell region. Two classes of lymphocyte video database provided by the cooperation hospital, Beijing You'an hospital, are used to verify the proposed approach. Experimental results show that it can quantitatively characterize lymphocyte and produce reasonably good agreement with subjective judgment by experts. Index Terms - Cell deformation characterization, image sequences analysis, cell image analysis, Fourier descriptors, quantitative characterization.