Hello,
I am trying to detect big black bars in pdf scans saved as jpgs. The big black bars are texts/passages that were blacked. To detect those in future scans I thought I can go with a Image Classification and use google/vit-base-patch16-224. Finally, the feature extractor just returns 1s as tensors which made me think if I can use VIT for that kind of problem. Does anyone has a recommendation? Or do I just have to adapt the feature extractor somehow? Thank you