WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation

Unofficial reimplementation of: [CVPR 2023] WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation [Paper]

Citation:

@InProceedings{Jeong_2023_CVPR,
    author    = {Jeong, Jongheon and Zou, Yang and Kim, Taewan and Zhang, Dongqing and Ravichandran, Avinash and Dabeer, Onkar},
    title     = {WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2023},
    pages     = {19606-19616}
}

Install Dependencies

TO BE UPDATED

Run Evaluation

To run evaluation, go to repo folder, modify eval.py on ds, param.py for directory, device, etc

python eval.py <shots> -mode

to run evaluation on , mode = AC or AS for anomaly classification or segmentation

Get Image segmentation result visually

Modify data directory in param.py, modify directory, shot, object name in segmentation.py to specify where to save result, number of shot, object

python segmentation.py

get result in ur specified directory

Data

Produce promissing results on MVtecAD
Also available for VisA, (with limitations)

Acknowledgement

Many codes borrowed from OpenCLIP and caoyunkang, your works helped me a lot during the process!