Optimized volumetric visual hull reconstruction method based on CUDA

Volumetric visual hull reconstruction is easy to be implemented without complicated geometrical computation, but its accuracy and efficiency are still unsatisfactory. An optimized volumetric visual hull reconstruction method is proposed based on CUDA. The voxel intersection judgment procedure and the isosurface extraction procedure are parallelized and implemented using multiple threads of CUDA to accelerate the shape modeling step. During triangulation step, marching cubes algorithm is improved through exact intersection points searching and redundant edge computation elimination. The pixel shader of CUDA is used to carry out the hybrid weight texture mapping so as to enhance the quality of the reconstructed mesh model. Experimental results on multi-camera platform DreamWorld proved the effectiveness of our method.