Polarization-based Uncalibrated Photometric Stereo 2020


Target	Result

We developed a method to estimate 3D shape of opaque object. Input images with different light source direction is used. This method can estimate the shape under unknown light source directions. We use the constraints that the light intensity is constant and the shape at the object boundary is known. The ambiguity problem of uncalibrated photometric stereo is solved using these constraints, which results in unique shape. We also developed a method to improve the shape using polarization.

Uncalibrated photometric stereo refined by polarization angle: Daisuke Miyazaki, Shuhei Hashimoto; Optical Review
vol. 28, no. 1, pp. 119-133, 2021/1/16
https://www.springer.com/journal/10043; Paper: English
Hashimoto-OR2021.pdf
https://doi.org/10.1007/s10043-021-00640-0

Uncalibrated photometric stereo constrained by intrinsic reflectance image and shape from silhoutte: Shuhei Hashimoto, Daisuke Miyazaki, Shinsaku Hiura; 16th International Conference on Machine Vision Applications (MVA2019)
01-05, pp. 1-6, 2019/5/28 (2019/5/27-31)
National Olympics Memorial Youth Center, Tokyo, Japan; http://www.mva-org.jp/mva2019/; Paper: English, Oral presentation: Japanese
Hashimoto-MVA2019.pdf
https://doi.org/10.23919/MVA.2019.8758025

Back to publication page