Closed caihaocong closed 7 months ago
Maybe you can adjust "min_max_norm" and "gaussian_filter" in eval.py.
Maybe you can adjust "min_max_norm" and "gaussian_filter" in eval.py.
Yes, that is right. Thank you very much ! Next, we will try to accelerate the inference speed of the model.
Hello, thank you for your open-source code. I train the dataset of the MvTecAD. I get the good performance of the anomaly images. but when I test the normal images, the image score is low but some region of the heatmap is high. What is the relationship between the image score and heatmap? How can I to reduce high heatmap of the normal images?