https://github.com/bchao1/photometric-stereo

Photometric stereo algorithms in numpy and pytorch.
https://github.com/bchao1/photometric-stereo

Last synced: about 2 months ago
JSON representation

Photometric stereo algorithms in numpy and pytorch.

• Host: GitHub
• URL: https://github.com/bchao1/photometric-stereo
• Owner: bchao1
• Created: 2022-04-11T07:52:09.000Z (over 3 years ago)
• Default Branch: master
• Last Pushed: 2022-04-18T03:29:41.000Z (over 3 years ago)
• Last Synced: 2025-06-03T02:29:31.324Z (5 months ago)
• Language: MATLAB
• Homepage:
• Size: 43.4 MB
• Stars: 10
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Photometric Stereo

Estimating albedo, surface normal, depth from multiple illuminated images.

# About

This repo implements some photometric stereo algorithms. The algorithms can be found the following papers:

- Shape and Albedo from MUltiple Images using Integrability (Yuille and Snow)

- Resolving the Generalized Bas-Relief Ambiguity by Entropy Minimization (Alldrin et al.)

- A Closed-Form Solution to Uncalibrated Photometric Stereo via Diffuse Maxima (Favaro et al.)

- Self-calibrating Photometric Stereo (Shi et al.)

# Dataset

[Harvard Photometric Stereo Dataset](http://vision.seas.harvard.edu/qsfs/Data.html)

# Results

## Albedo, normals, and depth estimation

![est](./results/result.gif)

## Resolving GBR ambiguity

|cat|women|frog|

|--|--|--|

|![gbr](./results/compare_cat.gif)|![gbr](./results/compare_women.gif)|![gbr](./results/compare_frog.gif)|

## CPU and GPU versions

Both Numpy and Pytorch implementations are provided. The code is tested on a `Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz` CPU and a `GeForce GTX 1080 Ti` GPU.

|Dataset|Numpy|Pytorch|

|--|--|--|

|cat|93.46s|10.61s|

|women|71.58s|9.54s|

|frog|103.13s|10.94s|

# Todo

- Coarse to fine refinement

- Paper: Reflections on the Generalized Bas-Relief Ambiguity