Open bdaad opened 1 year ago
I solved it problem. I will share the solution.
file : /patchcore/patchcore-inspection/build/lib/patchcore/utils.py 〔Before code〕 axes[2].imshow(segmentation)
〔After code〕 axes[2].imshow(segmentation, vmin=0, vmax=1,cmap=plt.cm.jet)
*Roughly line 70.
Please,I want to predict point of abnormality. How should I modify the code to output a prediction graph
I tried to predict point of abnormality. But, Outputs a relative evaluation of each image. So, No color difference between good image heat map and bad image heat map.
[ex.] Output good image![cable_test_good_002 (1)](https://user-images.githubusercontent.com/78455651/225323683-8a98f047-4afe-41eb-a5af-931ae9fa60c9.png)
Output bad image![cable_test_cut_outer_insulation_009 (1)](https://user-images.githubusercontent.com/78455651/225323647-3eba2a14-1568-4ad3-ab14-5b921a9a29b5.png)
Please tell me how to change relative evaluation into absolute evaluation. For [ex.] image, I want good image to be blue and bad point of bad image to be yellow.(otherwise blue)