SimpleNet

SimpleNet: A Simple Network for Image Anomaly Detection and Localization

Note

• The default dataset is metal_nut from MVTec

Installation

conda create -n PyTorch python=3.8
conda activate PyTorch
conda install pytorch torchvision torchaudio cudatoolkit=10.2 -c pytorch-lts
pip install opencv-python==4.5.5.64
pip install scikit-learn
pip install scipy
pip install tqdm

Train

• Configure your dataset path in main.py for training
• Download pretrained weights and place it under weights folder, see Pretrained weights
• Run python main.py --train for training

Test

• Configure your dataset path in main.py for testing
• Run python main.py --test for testing

Results

| Backbone | Epochs | F1 (%) | Accuracy (%) | AUROC (%) | Pretrained weights | |:--------:|:------:|-------:|-------------:|----------:|------------------------------------------------------------------------------------------:| | ResNet18 | 150 | 97.4 | 95.7 | 99.1 | model | | ReNet34 | 150 | 97.4 | 95.7 | 99.3 | model | | ReNet50 | 150 | 100 | 100 | 100 | model | | ReNet101 | 150 | - | - | - | model | | ReNet152 | 150 | - | - | - | model |

Reference

• https://github.com/DonaldRR/SimpleNet