Threshold value for classifying image as anomaly or not. and add a method to increase the accuracy by retraining on the wrong predictions.

[shrinand1996]

Is your feature request related to a problem? Please describe.
I tried the library and trained it on the leather dataset. I played around and tried to introduce a threshold value based upon score.
The model failed to create a clear "boundary" for the defective and non defective samples.

Describe the solution you'd like
To overcome this problem, it will be better if we can re train the model based upon the wrong predictions.

Describe alternatives you've considered

• A clear and concise description of any alternative solutions or features you've considered.

Additional context

• Add any other context or screenshots about the feature request here.