This Project implements an automated capsule endoscopy image classification system using the InceptionV3 deep learning model. It classifies gastrointestinal images into ulcers, polyps, bleeding, normal tissue using transfer learning, data augmentation, and CNN-based feature extraction to improve diagnostic accuracy and reduce manual review effort.