Diagnosis of leukemia cell from microscope images with image processing methods

In this thesis, computer vision and image processing tools were used to apply various algorithms to diagnose and detect leukemia cells. The one of the most dangerous disease is the leukaemia at nowadays. According to new scientific research, one million people die annually because of this disease. Early diagnosis is a very important factor for the treatment of leukaemia that's why research on the diagnosis of this problem has spread to other areas outside the biology. In this study, an efficient image processing algorithm is designed to recognize acute lymphocyte leukemia (ALL) cells, which are more common in children, have a high chance of treatment and can result in death if untreated. SVM (Support Vector Machine) is used as the method and data is pre-processed with wavelet transform. The results were statistically analyzed with the help of confusion matrix. The rate of success was found to be 95,700% for cancer data and 96,466% for non-cancer data.