Blood cell image localized segmentation combining mean-shift and ELM

The paper locate leukocyte nucleus using mean shift based on distribution characteristic of leukocyte nucleus in color space , then mark each leukocyte nucleus using Marking Algorithm based on matrix. Then the region of nucleus would be inflated in image in order to get a part of color information of cytoplasm around nucleus, by using entropy. A part of color of each cytoplasm and nucleus would comprise the positive training subset, and the rest do the negative training subset , A two-class ELM could be trained with the training set and generate a classification model .So several local models of ELM can be produced in the image in order to extract leukocyte one by one. The proposed algorithm need not tune the parameters, and it is better in segmentation effect, compared with SVM algorithm and the algorithm which extracts leukocytes entirely.