• Asha Gowri M
• Ashwija
• Deepika
• Lavanya Moger
• Mithun A

This study presents an advanced AI-driven framework for generating precise textual descriptions from medical endoscopy images, leveraging dual deep learning architectures: a vision transformer (ViT) for hierarchical feature extraction and a reinforced transformer decoder with cross-modal attention for clinically coherent report generation. The system addresses critical gaps in automated GI lesion documentation by integrating spatial-contextual awareness with domain-specific language modeling, outperforming existing CNN-LSTM hybrids in both accuracy and clinical relevance.

Endoscopy Image Captioning, Vision-Language Transformers, Uncertainty Quantification, Medical Report Generation, Few-Shot Learning, Clinical Decision Support.

"GASTROINTESTINAL DISORDER DETECTION USING ENDOSCOPY IMAGES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 7, page no.190-194, July-2025, Available :http://www.jetir.org/papers/JETIRGX06035.pdf

"GASTROINTESTINAL DISORDER DETECTION USING ENDOSCOPY IMAGES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 7, page no. pp190-194, July-2025, Available at : http://www.jetir.org/papers/JETIRGX06035.pdf