INTELCAPE: A Deep Learning-Powered System for Automated, High-Accuracy Crohn's Disease Diagnosis via Capsule Endoscopy - PubMed
5 hours ago
- #Capsule Endoscopy
- #Crohn's disease
- #Artificial Intelligence
- INTELCAPE is a deep learning-powered system for automated Crohn's disease (CD) diagnosis via capsule endoscopy (CE).
- The system integrates ResNet, Transformer, and EfficientNet architectures for hierarchical processing of CE videos.
- INTELCAPE achieved high accuracy in small-intestine segmentation (IoU scores up to 96.87%) and lesion detection (AUC up to 0.993).
- For CD diagnosis, INTELCAPE demonstrated robust generalizability with AUCs of 0.982 and 0.984, comparable to specialists but 10-fold faster.
- The system improved doctors' diagnostic accuracy from 76.7% to 94.8% and reduced interpretation time from 67.9 to 22.5 minutes.
- INTELCAPE is particularly beneficial for less-experienced clinicians, enhancing both accuracy and efficiency in CD diagnosis.