search

phuang17 / DeepMVS

DeepMVS: Learning Multi-View Stereopsis
https://phuang17.github.io/DeepMVS/index.html
BSD 2-Clause "Simplified" License
331 stars 85 forks source link

About calculate photometric errors #3

Closed tiffany61706 closed 6 years ago

tiffany61706 commented 6 years ago

Dear authors, Thanks for your work and open source of DeepMVS! I was wondering that in your paper you mentioned about how to calculate photometric and geometric errors, but I couldn't find it in your code. How did you calculate both of them? Thanks

phuang17 commented 6 years ago

At current stage, I have not provided the codes for quantitative evaluation yet. Nevertheless, thanks for the reminder and I think it makes sense to add those codes. I will probably add them in mid-June.

tiffany61706 commented 6 years ago

Thank you for your prompt reply. I am looking forward to your update. By the way, I have downloaded the ground truth of ETH3D dataset from your website. In your test code, you has ranged the disparity value into [0,1] to form into a output disparity image. And I use the same way to output the result as well and my image looks like grayscale image. Have you done any modification to let your images look like rgb image? Or you rearranges the value of disparity? Thanks again

phuang17 commented 6 years ago

I used the "viridis" colormap from matplotlib for disparity map visualization. Please refer to: https://matplotlib.org/examples/color/colormaps_reference.html

phuang17 commented 5 years ago

@tiffany61706 The evaluation codes have been added. Please open another issue or file a PR if there is any problem. Thanks!