layumi / HQ-Market

Market-1501 dataset with super-resolution quality

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Market-HQ

We enlarge the Market-1501 from 128x64 to 512x256 for 3D Mesh Generation. Here we provide a brief script for quick converting.

Download the super-resolution dataset:

  • Market-HQ

Download the processed data from Google Drive. Or use the gdrive tool to download via command line:

gdrive download 10gNi3n8Iny4O4MOZRs5dNFICfj8ri9eW

Generat data by yourself:

Dependency:

  1. de-jpeg compression (We do not use it in practice, but it works in some general cases.) https://github.com/victorvde/jpeg2png

  2. super-resolution https://github.com/xinntao/Real-ESRGAN

Generation:

Please modify the dataset root before running.

python prepare-hq.py 

Citation

Please cite this paper if it helps your research:

@inproceedings{zheng2015scalable,
  title={Scalable Person Re-identification: A Benchmark},
  author={Zheng, Liang and Shen, Liyue and Tian, Lu and Wang, Shengjin and Wang, Jingdong and Tian, Qi},
  booktitle={Computer Vision, IEEE International Conference on},
  year={2015}
}

@InProceedings{wang2021realesrgan,
    author    = {Xintao Wang and Liangbin Xie and Chao Dong and Ying Shan},
    title     = {Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data},
    booktitle = {International Conference on Computer Vision Workshops (ICCVW)},
    date      = {2021}
}

@article{zheng20223d,
  title={3D Magic Mirror: Clothing Reconstruction from a Single Image via a Causal Perspective},
  author={Zheng, Zhedong and Zhu, Jiayin and Ji, Wei and Yang, Yi and Chua, Tat-Seng},
  journal={arXiv preprint arXiv:2204.13096},
  year={2022}
}

About

Market-1501 dataset with super-resolution quality

License:MIT License


Languages

Language:Python 100.0%