COde: A benchmark multimodal oro-dental dataset for large vision-language models

This dataset is associated with our paper, currently under peer review, titled:
"A benchmark multimodal oro-dental dataset for large vision-language models."

The dataset is also available on GitHub: https://github.com/zirak-ai/COde

Introduction

Comprehensive multimodal dataset for AI dentistry, containing 8,775 dental checkups from 4,800 patients collected between 2018 and 2025. The dataset spans ages 10–90 and includes 50,000 intraoral photographs, 8,056 radiographs, and detailed clinical text, including diagnoses, treatment plans, and follow-up notes. All records were collected under ethical guidelines and annotated for benchmarking. We demonstrate the dataset’s value by fine-tuning Qwen-VL 3B and 7B on anomaly classification and diagnostic report generation. Fine-tuned models significantly outperform their base versions and GPT-4o, highlighting the dataset as a valuable public resource for future oral healthcare research across diverse clinical settings and multimodal learning tasks.