MSFR

KO | EN

gitlite — search

#react #ai #go #golang #machine-learning #rust #deep-learning #linux #nodejs #java #awesome #cli #mcp #open-source

(★ 19)

This is official implementation of Multi-scale feature reconstruction network for industrial anomaly detection

MSFR Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

**[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/multi-scale-feature-reconstruction-network/anomaly-detection-on-aebad-s)](https://paperswithcode.com/sota/anomaly-detection-on-aebad-s?p=multi-scale-feature-reconstruction-network)**
**[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/multi-scale-feature-reconstruction-network/unsupervised-anomaly-detection-on-aebad-s)](https://paperswithcode.com/sota/unsupervised-anomaly-detection-on-aebad-s?p=multi-scale-feature-reconstruction-network)**
**[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/multi-scale-feature-reconstruction-network/anomaly-detection-on-mvtecad-10)](https://paperswithcode.com/sota/anomaly-detection-on-mvtecad-10?p=multi-scale-feature-reconstruction-network)**
**[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/multi-scale-feature-reconstruction-network/unsupervised-anomaly-detection-on-mvtecad-1)](https://paperswithcode.com/sota/unsupervised-anomaly-detection-on-mvtecad-1?p=multi-scale-feature-reconstruction-network)**
## Multi-scale feature reconstruction network for industrial anomaly detection

This is an official PyTorch implementation of the paper [Multi-scale feature reconstruction network for industrial anomaly detection](https://www.sciencedirect.com/science/article/pii/S095070512401284X).
```
@article{iqbal2024multi,
  title={Multi-scale feature reconstruction network for industrial anomaly detection},
  author={Iqbal, Ehtesham and Khan, Samee Ullah and Javed, Sajid and Moyo, Brain and Zweiri, Yahya and Abdulrahman, Yusra},
  journal={Knowledge-Based Systems},
  pages={112650},
  year={2024},
  publisher={Elsevier}
}
```

### Datasets

Compared to existing datasets, AeBAD has the following two characteristics: 1.) The target samples are not aligned and at different sacles. 2.) There is a domain shift between the distribution of normal samples in the test set and the training set, where the domain shifts are mainly caused by the changes in illumination and view.

**Download dataset at [here](https://drive.google.com/file/d/14wkZAFFeudlg0NMFLsiGwS0E593b-lNo/view?usp=share_link) (Google Drive) or [here](https://cloud.189.cn/web/share?code=nYraE3uMRJn2) (access code: g4pr) (Tian Yi Yun Pan).**

* AeBAD-S

### Get Started

#### Pre-trained models

Download the pre-trained model of MAE (ViT-large) at [here](https://dl.fbaipublicfiles.com/mae/visualize/mae_visualize_vit_large.pth).

#### Dataset

**MVTec:**

**Create the MVTec dataset directory**. Download the MVTec-AD dataset from [here](https://www.mvtec.com/company/research/datasets/mvtec-ad). The MVTec dataset directory should be as follows.

**AeBAD:**

Download the AeBAD dataset from the above link. The AeBAD dataset directory should be as follows.

**Note that background, view and illumination in the train set is different from test. The background, view and illumination in test is unseen for the training set.**

#### Virtual Environment

Use the following commands:
```
pip install -r requirements.txt
```

#### Train and Test for MVTec, AeBAD

Train the model and evaluate it for each category or different domains. This will output the results (sample-level AUROC, pixel-level AUROC and PRO) for each category. It will generate the visualization in the directory.

run the following code:

```
sh mvtec_run.sh
```

```
sh AeBAD_S_run.sh
```

TRAIN.MSFR.model_chkpt in MSFR.yaml is the path of above download model. TRAIN.dataset_path (TEST.dataset_path) is the path of data.
Set Test.save_segmentation_images as True or False to save processed image.

## Acknowledgement
We acknowledge the excellent implementation from [MAE](https://github.com/facebookresearch/mae), [ViTDet](https://github.com/facebookresearch/detectron2/tree/main/projects/ViTDet).

## License
The data is released under the CC BY 4.0 license.

MSFR

MSFR Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

MSFR

MSFR Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README