Image Classification of Abnormal Red Blood Cells using Decision Tree Algorithm

This study presents a comprehensive web-based system for the automated classification of abnormal red blood cells (RBCs) using machine learning and image processing techniques.The system employs a decision tree classifier to identify and categorize eight types of abnormal RBCs: Degmacyte, Drepanocyte, Dacrocyte, Acanthocyte, Echinocyte, Elliptocyte, Spherocyte, and Codocyte.Developed using Flask, the application integrates computer vision methods such as Sobel edge detection, morphological analysis, and feature extraction.Key features include intensity histograms, texture properties from the Gray-Level Co-occurrence Matrix (GLCM), and geometric features like area, perimeter, circularity, and solidity.OpenCV is used for preprocessing tasks including grayscale conversion, resizing, and contour detection.The system offers a complete workflow from image upload to classification, supporting user authentication, training data management, model training, and prediction.Users can upload images, train decision tree models, and classify new microscopic images.Visualization tools such as decision tree diagrams, confusion matrices, and feature importance plots enhance interpretability.Performance evaluation confirms the system's effectiveness in differentiating morphologically similar RBC types using shape and texture analysis.The web-based interface provides an accessible, reliable diagnostic support tool for medical professionals, reducing manual errors in traditional RBC classification in clinical laboratories.