There is a problem with the Extract Mouth Batch.py script, Instead of cropping the mouth region it crops the upper eye.
I identified that this has something to do with the Width and the Height set for it. When you swap the values for the 2 variables it does provide a mouth render but in a skewed way. It would be helpful if you could provide examples of how the images should be output and correct the requirements for the training
There is a problem with the Extract Mouth Batch.py script, Instead of cropping the mouth region it crops the upper eye. I identified that this has something to do with the Width and the Height set for it. When you swap the values for the 2 variables it does provide a mouth render but in a skewed way. It would be helpful if you could provide examples of how the images should be output and correct the requirements for the training
Before Swapping the Values
. After swapping the values of width and height