Nowadays the primary illness of death in many populations is leukemia. The prognosis and possibility of recovery are significantly affected by the disease's early identification and diagnosis. Nowadays, blood related problems are diagnosed by visual evaluation of microscopic images, which involves looking at changes in the photos’ texture, shape, color, and statistical analysis. This system provides an early attempt to identify if a person has been affected by leukemia using blood sample photos. The ability to detect and diagnose leukemia at an earlier stage is made possible via image analysis. Due to their low cost and lack of need for expensive testing and lab equipment, images are used. In this study, generative adversial network (GAN) model is utilized to identify leukemia cells among the healthy blood cells. These damaged white blood cells can adhere to soft tissue called bone stem that can be found inside most bones. The abnormal white blood cells continue to grow uncontrollably in the bone marrow. Segmentation of the blood cells is done with k-means clustering.