FEATURE-BASED MACHINE LEARNING FRAMEWORK FOR THE DETECTION AND CLASSIFICATION OF ACUTE LYMPHOBLASTIC LEUKEMIA

Abstract

Early detection of Acute Lymphoblastic Leukemia (ALL) is essential for effective diagnosis and timely treatment. This study presents an automated framework for leukemia classification using microscopic blood smear images combined with machine learning techniques. Spatial features, including shape and texture, along with color-based descriptors, were extracted from microscopic images and analyzed both individually and in fused combinations. Several classifiers, such as k-Nearest Neighbor, Support Vector Machine, and Decision Tree, were employed to evaluate model performance. The results revealed that individual feature sets provided moderate accuracy, while the fusion of all three feature types substantially improved classification performance. Among the tested models, the Support Vector Machine achieved the highest accuracy of 100% using the combined feature set, reflecting its strong generalization capability. The proposed approach offers a cost-effective, accurate, and scalable diagnostic solution, highlighting its potential for integration into clinical workflows and automated leukemia screening in hematological analysis.

Keywords : Microscopic Blood Smear Images, Spatial Features, Acute Lymphoblastic Leukemia, Color Features, Segmentation, Machine Learning, Prediction.

https://doi.org/10.5281/zenodo.17403387