junyanz / pytorch-CycleGAN-and-pix2pix

Image-to-Image Translation in PyTorch
Other
23.09k stars 6.32k forks source link

Add ROI specific loss function to generator #1637

Open Diogo-Valente2111 opened 7 months ago

Diogo-Valente2111 commented 7 months ago

I would like to add an additional L1 loss function directly on the region of interest (bounding boxes) during training (Pix2pix) to try to improve clarity for object detection. Is it expected to yield better results? What are the steps that I have to follow to accomplish this?