Robust multi-view stereo synthesized by various parameters model

In this paper, we have developed a novel and robust framework of combining a matrix splitting with multi-view stereo reconstructions to separate reconstruction inaccuracies from a various parameters model for high-accuracy multi-view stereo reconstruction. Instead of performing the fixed parameters reconstruction procedure, we apply the variational based 3D reconstruction algorithm multi-times with various parameters to derive a set of hypothetic 3D models, and then synthesized the final result by formulating the problem as a low-rank matrix splitting problem. Benefited from the matrix splitting formulation, the outliers and bad matches, which are treated as the noise in the synthesized model, are effectively removed and thus lead to a 3D reconstruction with higher accuracy than the existing fixed parameters reconstructions. Constrained convex optimization is introduced for matrix splitting with an accelerated proximal gradient (APG) algorithm integrated for fast convergence. Both the experiments on the Middlebury and real-world data sets have demonstrated the effectiveness of the proposed method.