Detection of Cancer in Human Blood Sample using Machine Learning

Abstract

The process of identifying blood problems involves a human being looking at a blood sample under a microscope with their unaided eyes. In this study, a computerized method was created to aid doctors in recognizing various forms of leukaemia. Initial segmentation is performed using K-Mean clustering once the RGB image has been transformed to L*a*b color space. The properties of this clustered image are extracted and divided into various forms of leukaemia. This method is used to recognize the illnesses and provide an early diagnosis. Since images are inexpensive and don't require any expensive testing or lab equipment, they are used as inputs. In order to investigate any changes in colour, texture, geometry, and statistical analysis of the images, this research will make use of features in microscopic photographs. Proposed method will feed the changes discovered in these features into our classifier. Since images are inexpensive and don't require expensive testing or lab equipment, they are used. Leukemia, a disease of white blood cells, will be the system's main focus. The system will make advantage of microscopic picture attributes to analyses statistical changes in texture, geometry, and color.