identifies a photo in a picture and crops the background away
- Label some sample images of photos against the dark background (used CVAT)
- Train a yolo model to recognize photos against the dark background
- Run inference on full dataset
- Crop to the bounding box identified for each image
- Write output images to out an output folder