Novel depth-based blending technique for improved virtual view synthesis

The paper presents a new depth-based view blending technique that avoids the problem of different fields of view corresponding to the input views that are used for synthesis of a virtual view. The idea consists in using the depth associated with the input views in order to increase the quality of the finally blended view. The experiments, performed on high quality multi-view test sequences, show that the proposed method significantly improves the quality of synthesized views in systems with non-linearly arranged cameras.