depth from polarisation

[faezeh-mortazavi]

Dear prof. William A. P. Smith @waps101
Thank you for sharing the codes. I have a question about it: I saw the results of surface normals in your amazing paper. How can we estimate surface normals from these codes?

[waps101]

I recommend using the following function from another of my repositories:

https://github.com/waps101/polarisation_optimisation/blob/master/utils/depthnormals_mask.m

This computes a normal map from a height map via finite difference derivatives. Apply this to the height map you get from shape-from-polarisation.