A New Improved Depth Estimation Algorithm Based on Gradient Feature for Preserving Consistency among Views

In this paper we introduce a more reliable depth estimation algorithm, which is effective in improving the depth maps. With the commercial development of three-dimensional television (3DTV), the corresponding three-dimensional productions are becoming more popular. Especially, the depth estimation is one of the key technologies of multi-view video (MVV) and free viewpoint video (FVV) systems, but how to get accurate depth information is becoming the key issue. Based on the existing depth estimation reference algorithm (DERS), the proposed algorithm takes the improvements of the data term of the global algorithm. As experiment results show, our method not only improves the objective quality of synthesized virtual images in terms of PSNR, but also makes the subjective visual effect better.