Error in the code

[agoodge]

Hi, Lines 841-846 in multiple_features.py reads:

        self.xyz_mean = torch.mean(self.patch_xyz_lib)
        self.xyz_std = torch.std(self.patch_rgb_lib)
        self.rgb_mean = torch.mean(self.patch_xyz_lib)
        self.rgb_std = torch.std(self.patch_rgb_lib)
        self.fusion_mean = torch.mean(self.patch_xyz_lib)
        self.fusion_std = torch.std(self.patch_rgb_lib)

Is this correct? It looks like an error.

Thanks

[leefty-1024]

I have same question. But the result is bad if I fix it

[nomewang]

This code mainly does scaling on the values of the data, because there is some instability in the image edge values of the XYZ data, so we use a uniform scaling factor and mean value. Maintaining the way the code is written works better.