Classification and automatic segmentation for Detection Leukemia

The proposed work is to present a methodology for segmenting the nuclei of white blood cells based on Gram Schmidt orthogonalization & Sparse Representation for white blood cells classification. The differential counting of white blood cells reveals invaluable information to hematologist. These informations are very useful to hematologist for diagnosis and treatment of many diseases. The nucleus of white blood cells has the most information about type of white blood cells thus an accurate segmentation of white blood cell’s nucleus seems to be helpful for other stages of automatic recognition of white blood cells. The system will focus on features of white blood cells to detect Leukemia disease.The system will use features of microscopic images & examine changes on texture geometry color statistical. Changes in these features will be used as a classifier input.