Detection of Leukemia and Its Subtypes Using MATLAB: An RGB Color Code Analysis

Abstract

Objective: To develop and evaluate an automated technique for the detection and classification of leukemia subtypes using RGB color code analysis combined with image segmentation. Study Design: Experimental study. Place and Duration of Study: Lady Reading Hospital, Peshawar, from January 2023 to February 2025. Methods: An RGB-based color thresholding method was designed with a 3D median filter to isolate leukemic cells from stained blood smear images. The technique segmented out unwanted regions such as red blood cells and background artifacts while preserving diagnostic features of abnormal white blood cells. Color codes (RGB and hexadecimal values) were extracted from the segmented cells and stored in Excel format. Both automated and manual analysis of these color patterns were performed to distinguish between leukemia subtypes. The system was tested primarily on Acute Myeloid Leukemia (AML) and Acute Lymphocytic Leukemia (ALL) images, with further potential application to Chronic Myeloid Leukemia (CML) and Chronic Lymphocytic Leukemia (CLL). Results: The segmentation algorithm achieved high accuracy, with 97.63% for AML and 97.64% for ALL. Distinct color code patterns were successfully identified and associated with specific leukemia subtypes. The manual method also supported color recognition for subtype classification, enhancing diagnostic clarity. Conclusion: The proposed automated approach using RGB color code analysis and image segmentation provides a reliable, efficient, and scalable framework for leukemia detection. It enhances diagnostic accuracy and supports subtype differentiation, offering valuable support for clinical and computational hematopathology.

Keywords: Leukemia Detection; Acute Leukemia; Chronic Leukemia; Color Coding; Image Segmentation; RGB Thresholding.