A new strategy for urinary sediment segmentation based on wavelet, morphology and combination method

This paper presents a strategy for segmenting urinary sediment based on wavelet, morphology and combination method. Firstly, the wavelet transforms and morphology are used to get rid of the effect of the defocusing and get the subimages that include the particles. Then based on the characteristics of the subimages, edge detection and adaptive thresholding are employed adaptively. Finally, a simplified watershed algorithm for the overlapping particles is used. The experiment results show that the method can segment the defocusing urinary sediment images effectively and precisely.