Clustering-based match propagation for image-based rendering from multiple images

In this paper we propose a novel method for rendering a new view from multiple input images. A set of reliable pixel matches are initially selected and then propagated to neighboring pixels based on the piecewise smooth assumption in the spatial domain and the clustering-based photoconsistency constraint which is applied to handle pixel occlusion, partial pixel effects and lighting effects between images. Experimental results show that the proposed algorithm is promising and satisfactory.