Analysis and Recognition of Blood Cells to Detect Leukaemia Using Stack-Trained and Colour-Trained Algorithm
Abhishek Ghose Biswas, Ritika Sanwal, Sanvidha Haribhakta, Varsha Mittal, Pradeep Kumar Singh, Virendra Pal Singh · 2024
Leukemia is a lethal cancer condition that impacts people of every demographic, both kids and adolescents. It is a leading factor for mortality globally. It has a special relationship with White Blood Cells (WBC), followed by an increase in young lymphocytes, and triggers bone marrow and blood harm. As a result, timely and accurate cancer detection is essential for effective treatment and increased lifetime chances. At present, diagnosing the condition requires physical examination of blood specimens using microscopic visuals, which can be laborious and inaccurate. Under the microscope, cells with leukemia resemble normal cells, making diagnosis challenging. This paper introduces a novel stack and color-trained convolution neural network (CNN) for diagnosing leukemia blood cells using microscopic visuals of blood specimens. This study allowed us to autonomously recognize and categorize 7 types of blood cells, unlike previously reported techniques that often concentrated solely on a single kind of blood cell. This study concluded that by integrating statistical form and structural factors into the basic leukocyte numbers of traditional evaluation, CNN-based computerized blood cell assessment has a chance to simplify and enhance routine diagnoses.