Resolving Shape from Polarization Ambiguities Algorithm

This is an algorithm resolves the pi-ambiguity in shape from polarization (SfP) using the full polarimetric characterization of an object, in the form of a Mueller matrix. This algorithm simulates a monocular Mueller image of several surfaces and objects, and adds varying levels noise. There is a nonlinear optimization to estimate the scattering geometry from the simulated Mueller image. This inverse problem takes as inputs the camera and source locations, the camera properties, and the pBRDF model of the material the object is made of. Using the scattering geometry retrieved from the optimization, the algorithm calculates the ambiguous surface normals, as well as a depth estimate. This depth estimate is used to disambiguate the normals by providing a third surface normal estimate through the gradient of the depth. The depth estimate is highly susceptible to noise so image processing techniques are applied. The divergence of the ambiguous normals allows the object to be segmented into smooth areas of continuous curvature and polynomials are fitted to the depth to smooth it. Once the ambiguity is resolved, the mean angular errors of the normals and the percent of pixels disambiguated correctly are reported.

This dataset is associated with "Resolving Shape from Polarization Ambiguities" by McKenna & Kupinski submitted to IEEE Transactions on Image Processing.

For inquiries regarding the contents of this dataset, please contact the Corresponding Author listed in the README.txt file. Administrative inquiries (e.g., removal requests, trouble downloading, etc.) can be directed to data-management@arizona.edu