Unifying diffuse and specular reflections for the photometric stereo problem

R Mecca, Y Quéau - 2016 IEEE Winter Conference on …, 2016 - ieeexplore.ieee.org
2016 IEEE Winter Conference on Applications of Computer Vision (WACV), 2016ieeexplore.ieee.org
After thirty years of researching, the photometric stereo technique for 3D shape recovery still
does not provide reliable results if it is not constrained into very well-controlled scenarios. In
fact, dealing with realistic materials and lightings yields a non-linear bidirectional reflectance
distribution function which is primarily difficult to parametrize and then arduous to solve. With
the aim to let the photometric stereo approach face more realistic assumptions, in this work
we firstly introduce a unified irradiance equation describing both diffuse and specular …
After thirty years of researching, the photometric stereo technique for 3D shape recovery still does not provide reliable results if it is not constrained into very well-controlled scenarios. In fact, dealing with realistic materials and lightings yields a non-linear bidirectional reflectance distribution function which is primarily difficult to parametrize and then arduous to solve. With the aim to let the photometric stereo approach face more realistic assumptions, in this work we firstly introduce a unified irradiance equation describing both diffuse and specular reflection components in a general lighting setting. After that, we define a new equation we call unifying due to its basic features modeling the photometric stereo problem for heterogeneous materials. It is provided by making the ratio of irradiance equations holding both diffuse and specular reflections as well as non-linear light propagation features simultaneously. Performing a wide range of experiments, we show that this new approach overcomes state-of-the-art since it leads to a system of unifying equations which can be solved in a very robust manner using an efficient variational approach.
ieeexplore.ieee.org
以上显示的是最相近的搜索结果。 查看全部搜索结果