Adaptive Lagrangian Multiplier derivation model for depth map coding

In stereoscopic video coding, vast and varied coding modes are utilized to adapt various video content. In order to select the optimal coding parameter set for coding unit, rate-distortion optimization (RDO) criterion is employed as a critical technology. It has been found that the Lagrangian Multiplier of RDO criterion for depth map coding has an impact on coding performance of stereoscopic video. Therefore, the Lagrangian Multiplier should be determined properly. In this paper, we first provide an analysis of the factors that could affect the Lagrangian Multiplier. Then, an adaptive Lagrangian Multiplier derivation model for depth map coding in 3D-HEVC is established by involving these affect factors. At last, the optimal coding parameter set can be obtained by the adjusted RDO criterion through the derived Lagrangian Multiplier. Experimental results demonstrate that the proposed adaptive Lagrangian Multiplier derivation model can lead to a favorable coding performance of stereoscopic video. An average 3.13% bitrate saving can be achieved compared with the HTM 16.0 reference software. Moreover, compared to two other state-of-art methods, namely the Yang's method and the Huo's method, the proposed model can achieve 0.21% and 0.11% bitrate savings, respectively.