遇见数据集

RDG_Dataset

收藏
魔搭社区2026-07-13 更新2026-07-15 收录
官方服务:

资源简介:

### Data overview We use the Cycles engine to render the scene model and generate the dataset. - The collected dataset covers different scenarios such as complex textures and geometries, glossy reflections, and fast-moving objects. - For each frame, we render the paired LR-HR data with a size of 270\*480 pixels and 1080\*1920 pixels, respectively. - Each frame contains BRDF, normal, depth and motion vector - The G-buffer components for each frame include normal, BRDF, depth, and motion vector. ### Data structure The dataset has the following structure: ``` |--dataset |--training data |-- HR |--video 0 |--BRDF |--frame 0 : |--frame 99 |--Depth |--frame 0 : |--frame 99 |--Emit |--frame 0 : |--frame 99 |--Image |--frame 0 : |--frame 99 |--Motion |--frame 0 : |--frame 99 |--Normal |--frame 0 : |--frame 99 |--video 1 : |--video 199 |-- x4LR |--video 0 |--BRDF |--frame 0 : |--frame 99 |--Depth |--frame 0 : |--frame 99 |--Emit |--frame 0 : |--frame 99 |--Image |--frame 0 : |--frame 99 |--Motion |--frame 0 : |--frame 99 |--Normal |--frame 0 : |--frame 99 |--video 1 : |--video 199 |--test data |-- HR |--video 0 |--BRDF |--frame 0 : |--frame 199 |--Depth |--frame 0 : |--frame 199 |--Emit |--frame 0 : |--frame 199 |--Image |--frame 0 : |--frame 199 |--Motion |--frame 0 : |--frame 199 |--Normal |--frame 0 : |--frame 199 |--video 2 : |--video 7 |-- x4LR |--video 0 |--BRDF |--frame 0 : |--frame 199 |--Depth |--frame 0 : |--frame 199 |--Emit |--frame 0 : |--frame 199 |--Image |--frame 0 : |--frame 199 |--Motion |--frame 0 : |--frame 199 |--Normal |--frame 0 : |--frame 199 |--video 1 : |--video 7 ``` ### Clone with HTTP ``` git clone https://www.modelscope.cn/datasets/Sunny21/RDG_Dataset.git ``` ### Contact If you have any questions, please feel free to reach us out at cs.longsun@gmail.com ### Citation If you find our work helpful for your research, please consider giving a star ⭐ and citation 📝 ```bibtex @InProceedings{RDG, title={Efficient Video Super-Resolution for Real-time Rendering with Decoupled G-buffer Guidance}, author={Zheng, Mingjun and Sun, Long and Dong, Jiangxin and Pan, Jinshan}, booktitle={CVPR}, year={2025} }

提供机构:
maas
创建时间:
2026-05-21
二维码
社区交流群
二维码
科研交流群
商业服务