About ground-truth of the 90 real training data

Vandermode / ERRNet

Single Image Reflection Removal Exploiting Misaligned Training Data and Network Enhancements (CVPR 2019)

MIT License

247 stars 52 forks source link

Closed JiaxiongQ closed 5 years ago

JiaxiongQ commented 5 years ago

Should the real training data be normalized to range [0,1] during training? I find your training code doesn't do that.

Vandermode commented 5 years ago

Simply dividing 255 is what you want.