search

JosephChenHub / DPANet

DPANet : Depth Potentiality-Aware Gated Attention Network for RGB-D Salient Object Detection
MIT License
30 stars 10 forks source link

Normalize of the depth image #6

Closed Icecream-blue-sky closed 2 years ago

Icecream-blue-sky commented 2 years ago

When normalize the depth image, why set self.d_mean = 116.09, self.d_std = 56.61? How do you caculate this? Thanks!

JosephChenHub commented 2 years ago

When normalize the depth image, why set self.d_mean = 116.09, self.d_std = 56.61? How do you caculate this? Thanks!

Hi, you can find that we only normalize the depth map by dividing 255 in the definition of Normalize, that means the self.d_mean and self.d_std are not used.