Ulcer Detection and Classification into Normal, AVM, and Ulcer Conditions Using MobileNetV3

Ulcer detection in medical imaging is a critical task for early diagnosis and effective treatment planning, particularly in gastrointestinal examinations. Conventional diagnostic approaches rely heavily on clinicians’ expertise and manual inspection of endoscopic images, which is time-consuming and prone to human error. To address these limitations, this work proposes an automated deep learning- based ulcer detection and classification system using the MobileNetV3 architecture. The model classifies medical images into three categories: Normal, Arteriovenous Malformation (AVM), and Ulcer. Preprocessing techniques such as normalization and data augmentation are applied to enhance image quality and improve feature extraction. MobileNetV3 is selected due to its lightweight architecture and efficiency. Experimental results demonstrate high accuracy, precision, recall, and F1-score. The proposed system reduces diagnostic workload and enables real-time deployment in resource-constrained healthcare environments.