SofiaOrtegat's shadingShapeOptimiser

Shape from Shading Optimiser

The primary methodologies and schema are outlined in Note.pdf within the Note folder.

Sample outcomes are provided on the wiki page.

The crux of the algorithm is inspired by Tai-Pang Wu and Chi-Keung Tang's work Dense Photometric Stereo Using a Mirror Sphere and Graph Cut, available in the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (2005). Graph Cut Fold-Fulkerson maxflow algorithm toolbox sourced from http://vision.ucla.edu/~brian/gcmex.html

Shape from shapelet reconstruction is referenced from Peter Kovesi's Shapelets Correlated with Surface Normals Produce Surfaces (IEEE International Conference on Computer Vision). The toolbox is available at http://www.csse.uwa.edu.au/~pk/Research/MatlabFns/

================

Getting Started: Open preProcessData.m, adjust the directory of input data, set correct image size, and copy the light source direction file into code folder.

Compile the mex file. Load lightvec.txt

surfaceNormal = initialNormal(VarName1, VarName2, VarName3); This step would generally require 1-2 minutes.

refineNormal = graphCutSurfaceNormal(surfaceNormal); This is a time-consuming step.

You can generate N-L image result by lightImg = diffuseLight(surfaceNormal); or lightImg = diffuseLight(refineNormal);

Use recsurf = shapeFromShapelets(surfaceNormal); or recsurf = shapeFromShapelets(refineNormal); to review 3D reconstruction outcome.