增强的深度建模模式Contour分割

Enhanced contour partition in depth modelling mode

在新一代高效三维视频编码标准(High Efficiency Video Coding for 3D,3D-HEVC)中,为了更加精确地表示深度图中的物体边界,利用非矩形块分割的深度建模模式(Depth modelling modes,DMMs)预测编码方法被引入,现有的DMMs模式包括锲形(Wedgelet)分割和轮廓(Contour)分割2种模式,其主要区别在于推导分割的方式不同。Contour分割利用已编码纹理块信息进行深度分区推导,然而这种分割方法仅仅利用了纹理与深度之间的结构相关性,而忽略了边界相邻块之间具有的边缘相似性,这潜在导致推导块分割不准确,进而影响编码效率。针对Contour分割模式导致的分割不准确问题,结合深度边界块之间的边缘相似信息,提出了一种增强的Contour分割方法。实验结果表明,与3D-HEVC的参考软件HTM11.0相比,在全帧内(AllIntra,AI)编码测试条件下,对于合成视点平均有近0.1%的BD-Rate节省。

In the new generation of video coding standard(High Efficiency Video Coding,HEVC)and 3Dvideo extension of HEVC standard known as 3D-HEVC,non-rectangular Wedgelet and Contour block partitions,namely Depth Modelling Modes(DMMs),are added to achieve better representation of edges in depth maps.The existing DMMs contain two modes,i.e.DMM1 and DMM4,corresponding to Wedgelet and Contour partition,respectively.The two partitions mainly differ in the way the partitioning is generated.For Contour partition,the corresponding coded texture information of current depth block is employed to predict the final block partition.However,it just utilizes the structural similarity between texture and depth,without considering the edge similarity between adjacent boundary blocks within the same depth map.It may lead to inaccurate block partition and affect the coding efficiency.To address the problem above,an enhanced Contour partition generation method is proposed to make full use of the edge similarity between adjacent boundary blocks.The experimental results demonstrates nearly averaged 0.1% BD-rate saving can be achieved for synthesized views under all intra case,compared with the reference software HTM 11.0.