Design a new algorithm to count white blood cells for classification leukemic blood image using machine vision system

White blood cells protect the immune system against viruses and bacteria. Data extraction from white blood cells may cause problems such as loosing form, dimensions and edges. In this study, a complete and automatic method to identify and classify white blood cells using microscopic images has been presented. In the proposed method, in the first step, white blood cells are identified using color space conversion models. Then leukocytes group are separated using division of watershed conversion. In the next step image cleanup is done and all leukocytes available on the edge of images and abnormal components are removed. This is accomplished by cutting the image with the smallest rectangle that has connected components. The second level of division relates to the detection of the nucleus and cytoplasm. In the last step feature extraction is performed which causes the pathologists can have the best interpretation of them. All the above steps have been performed in MATLAB software. At the end, the proposed method was examined by a database belonging to Imam Reza (AS) hospital in Mashhad, consisting of 29 images of blood cells, and showed the accuracy of 93% in the detection of white blood cells.