Crohn’s Disease Lesion Classification
Developed an AI-based framework for classifying Crohn’s disease lesions from endoscopic images, combining deep-learning feature extraction with topology-aware analysis and test-time techniques.
- Worked with ResNet18, DenseNet121 and Vision Transformer architectures.
- Explored Topological Data Analysis (TDA), persistent homology and TTA.
- Compared deep-learning and feature-based machine-learning approaches.