search

MTLab / rsgunet_image_enhance

Champion solution of the PIRM2018 Challenge on Perceptual Image Enhancement on Smartphones (Track B: Image Enhancement)
http://openaccess.thecvf.com/content_ECCVW_2018/papers/11133/Huang_Range_Scaling_Global_U-Net_for_Perceptual_Image_Enhancement_on_Mobile_ECCVW_2018_paper.pdf
MIT License
120 stars 31 forks source link

UNet.py does not include pooling #7

Open captainst opened 4 years ago

captainst commented 4 years ago

Great job. Congrads.

One quick issue I found is this: Your paper states that you used the U-Net structure, which includes the max-pooling to make "down-sample". I found that your UNet.py does not include this max-pooling. Instead, a convolution with stride of 2 is used to strink the tiles in steps. Is there an explicit purpose of this modification?

Many thanks.