论文部分内容阅读
多视视频编码是实现自由视角电视和三维立体电视等应用的关键技术。为了有效地提高视间预测的效率,该文基于摄像机几何模型,提出了一种新的视间预测算法,其特点是利用摄像机参数和深度信息对视差矢量进行预测,并运用亮度补偿工具和亚像素插值技术减少视差补偿误差。将新的算法集成到目前最新的多视视频编码框架H.264/AVCM VC中后,实验结果表明,新的算法最高能节省32.82%的码率,编码增益最高可达2.067 dB。
Multi-video coding is the key technology to achieve free-view TV and 3D stereoscopic television applications. In order to effectively improve the efficiency of inter-view prediction, this paper proposes a new inter-view prediction algorithm based on camera geometry model, which is characterized by the use of camera parameters and depth information to predict the disparity vector, and the use of brightness compensation tools and sub- Pixel interpolation reduces parallax compensation errors. After the new algorithm is integrated into H.264 / AVCM VC, the latest multi-view video encoding framework, the experimental results show that the new algorithm can save up to 32.82% code rate and up to 2.067 dB coding gain.