基于时间下采样的深度图编码方法

In three-dimensional video(3DV) systems,depth maps are not used for visual enjoyment directly but for rendering virtual views.To improve the coding efficiency of 3DV systems while maintaining the qualities of synthesized virtual views, a temporal subsampling based coding method for depth maps is proposed. In the proposed method, a sequence of depth maps is first temporally down sampled and encoded. Then, at the decoder, the discarded frames in the depth map sequence are interpolated by a depth reconstructed method. Experimental results demonstrate that a maximum 0.245 dB and an average 0.075 dB BD PSNR gain could be achieved by the proposed method while maintaining the same coding bit rate.