Detection of Acute Lymphoblastic Leukemia using Novel Bone Marrow Image Segmentation

This paper presents an automated system for detecting Acute Lymphoblastic Leukemia (ALL) using bone marrow image segmentation and deep learning techniques. The system applies preprocessing methods such as grayscale conversion, contrast enhancement, and noise removal to improve image quality. Segmentation is performed using thresholding and morphological operations to extract relevant regions. A fine-tuned Xception model is used for classification into multiple leukemia cell types. Performance is evaluated using accuracy, precision, recall, and F1-score. The proposed approach improves diagnostic efficiency and reduces manual effort, making it suitable for clinical applications.