---
license: apache-2.0
pipeline_tag: image-text-to-text
library_name: transformers
---

# M3D-RAD Model
The official Model for the paper "[3D-RAD: A Comprehensive 3D Radiology Med-VQA Dataset with Multi-Temporal Analysis and Diverse Diagnostic Tasks](https://huggingface.co/papers/2506.11147)".

In our project, we collect a large-scale dataset designed to advance 3D Med-VQA using radiology CT scans, 3D-RAD, encompasses six diverse VQA tasks: anomaly detection (task 1), image observation (task 2), medical computation (task 3), existence detection (task 4), static temporal diagnosis (task 5), and longitudinal temporal diagnosis (task 6). 

![Main Figure](https://github.com/Tang-xiaoxiao/M3D-RAD/blob/main/Figures/main.png?raw=true)

## Code
You can find our code in [M3D-RAD_Code](https://github.com/Tang-xiaoxiao/M3D-RAD).

## 3D-RAD Dataset
You can find our dataset in [3D-RAD_Dataset](https://huggingface.co/datasets/Tang-xiaoxiao/3D-RAD).

## Model Links

| Model | Paper                                                        |
| ----- | ------------------------------------------------------------ |
| [RadFM](https://github.com/chaoyi-wu/RadFM) | Towards Generalist Foundation Model for Radiology by Leveraging Web-scale 2D&3D Medical Data | https://github.com/chaoyi-wu/RadFM |
| [M3D](https://github.com/BAAI-DCAI/M3D)   | M3D: Advancing 3D Medical Image Analysis with Multi-Modal Large Language Models |
| OmniV(not open) | OmniV-Med: Scaling Medical Vision-Language Model for Universal Visual Understanding |