search

JeffWang987 / MVSTER

[ECCV 2022] MVSTER: Epipolar Transformer for Efficient Multi-View Stereo
MIT License
183 stars 13 forks source link

A qustion about mono depth #7

Open Wuuu3511 opened 1 year ago

Wuuu3511 commented 1 year ago

thanks for your code！ I have a question about MVS4Net.py “line 109 outputs = self.mono_depth_decoder(outputs, depth_values[:,0], depth_values[:,1])” depth_values[:,-1] seems to d_max rather than depth_values[:,1]. Is there a problem with my understanding?

zhz120 commented 1 year ago

I suspect that the author is just looking for innovations. This monocular depth estimation has no effect on training at all, and the final loss L1 weight is 0

Wuuu3511 commented 1 year ago

I suspect that the author is just looking for innovations. This monocular depth estimation has no effect on training at all, and the final loss L1 weight is 0

thanks for your reply！ but in the paper , this loss is helping to learn more distinguishing Fpn features.I think it should be used while training