ZJU-Leaper Dataset

This dataset aims to set a benchmark suit for vision based fabric defect detection, including image dataset, evaluation protocol, baseline experiments.

Please refer to our website (backup mirror) for more information.

Dataset API

https://github.com/nico-zck/ZJU-Leaper-Dataset/tree/master/dataset_api

Baseline experiments

https://github.com/nico-zck/ZJU-Leaper-Dataset/tree/master/baselines

Related Paper

The paper titled "ZJU-Leaper: A Benchmark Dataset for Fabric Defect Detection and a Comparative Study" was accepted by IEEE Transactions on Artificial Intelligence.

@article{zju-leaper,
  author={Zhang, Chenkai and Feng, Shaozhe and Wang, Xulongqi and Wang, Yueming},
  journal={IEEE Transactions on Artificial Intelligence}, 
  title={ZJU-Leaper: A Benchmark Dataset for Fabric Defect Detection and a Comparative Study}, 
  year={2020},
  volume={1},
  number={3},
  pages={219-232},
  doi={10.1109/TAI.2021.3057027}
  issn={2691-4581},
  publisher={IEEE},
}

TODO List

• codes for dataset API
• documentation for dataset API
• codes for baseline experiments
• download link for the dataset