Author(s) :   Dr. K. Pandikumar¹, Yogesh Kumar C², Rishikesh J³, Sudharsan P⁴, Santhosh P⁵

Abstract : This study proposes a deep learning-based system utilizing MobileNet architectures for the detection and classification of blood cell cancer into four categories: Benign, Malignant Pre-B, Malignant Pro-B, and Malignant Early Pre-B. The MobileNet model achieved a high accuracy of 97%, demonstrating its effectiveness in diagnosing blood cell abnormalities. The system includes a user-friendly frontend that not only displays the predicted cancer class and confidence score but also provides detailed descriptions of each cancer stage along with treatment recommendations, facilitates clinical decision-making. Experimental results highlight the model’s strong precision and recall, validating its capability for early and accurate detection. Additionally, comparative analysis confirms MobileNet’s superior performance in both accuracy and computational efficiency over other deep learning architectures, making it an ideal solution for deployment in real-world and resource-constrained healthcare environments. This approach offers a robust and efficient tool to support hematologists and researchers, contributing to improved patient outcomes through early diagnosis.

Keywords: Blood cell cancer, deep learning, MobileNet, cancer classification, early detection, clinical decision support, lightweight model, medical AI.

DOI: Available on author(s) request

Pages : 53-58

*Authors are invited to submit papers through E-mail at editor.ijarcsms@gmail.com