Lightweight Framework for RBC Disease Detection

Blood-related cancers such as leukemia and myeloma are among the most life-threatening diseases, with high mortality rates worldwide. Timely and accurate detection plays a pivotal role in improving patient survival rates. However, traditional diagnostic procedures such as laboratory blood tests, chemical analysis, and microscopic evaluations are often time-consuming, expensive, and inaccessible in resource-limited settings. These limitations delay treatment initiation and increase patient stress. This paper introduces an automated Red Blood Cell (RBC) disease classification system that leverages machine learning and digital image processing to provide faster and more affordable diagnostics. The proposed framework involves four major stages: preprocessing, segmentation, feature extraction, and classification. Preprocessing enhances the quality of blood smear images, segmentation isolates RBCs from other cells, feature extraction reduces data dimensionality while preserving meaningful attributes, and classification is carried out using the K-Nearest Neighbors (KNN) algorithm. The system is implemented using Python, TensorFlow, Flask, and cloud-based tools with datasets obtained from Kaggle. Results indicate that the proposed system significantly reduces diagnostic time while maintaining reliable classification accuracy. Unlike deep learning-based approaches, which require extensive computational resources, the presented system is lightweight, cost-effective, and adaptable to diverse healthcare settings. This makes it particularly useful for hospitals in rural and developing regions.