Red Blood Cell Classification through Depth Map and Surface Feature

The shape of red blood cell contributes more to clinical diagnosis with respect to blood diseases. In this paper, a compound classification algorithm for red blood cell was proposed, in which the depth map and surface curvature incorporated to deal with different kinds of cell shape. The original image was preprocessed to denoise and the individual 3-D shape of each cell was obtained with shape from shading technique, which was resolved by linear approximation under the partial differential equation IRE. The next procedure was to construct a segmentation algorithm based on surface curvature by fitting smooth multi-scale surface. The number of different shape could be used to classification. It is experimentally shown that the proposed algorithm can segment cells with complex surface shape successfully.