XiaLiPKU / EMANet

The code for Expectation-Maximization Attention Networks for Semantic Segmentation (ICCV'2019 Oral)
https://xialipku.github.io/publication/expectation-maximization-attention-networks-for-semantic-segmentation/
GNU General Public License v3.0
680 stars 130 forks source link

Is this a bug or trick? image = (image - settings.MEAN) / settings.MEAN #7

Closed gasvn closed 5 years ago

gasvn commented 5 years ago

https://github.com/XiaLiPKU/EMANet/blob/9a492d8aaad297e15eac044b3bb9583e63ffa3a3/dataset.py#L19

I support this line should be

image = (image - settings.MEAN) / settings.STD.

Or is this line a trick?

XiaLiPKU commented 5 years ago

Oh, this is really a bug in the released code. I will fix it right now. Thanks!

XiaLiPKU commented 5 years ago

But it seems that this line doesn't influence the final performance so much. No matter what, thanks for your carefulness.