Abstract:

Concerning the problem that the computational complexity of Coding Unit (CU) quad-tree splitting algorithm for High Efficient Video Coding (HEVC) intra prediction is very high, a fast intra CU splitting algorithm was proposed, which can narrow depth range of CU splitting. Firstly, multiple texture features were achieved through new defined texture extraction algorithm. Secondly, in order to get a decision function, Support Vector Machine (SVM) was employed to train the feature parameters. At last, based on decision function, CU splitting depth range was narrowed by skipping unnecessary splitting and early terminating splitting. The experimental results show that compared with the reference model HM 12.0, the fast CU splitting algorithm provides about 43.23% time savings on average with 0.84% increasing on bit-rate which obviously improves the coding efficiency. Besides, the proposed algorithm is easy to be combined with other methods to further reduce the computational complexity for HEVC intra coding.

Key words: High Efficiency Video Coding (HEVC), intra prediction, Coding Unit (CU), quad-tree splitting, Support Vector Machine (SVM)

摘要：

针对高效率视频编码(HEVC)帧内预测过程中,编码单元四叉树划分算法计算复杂度极高的问题,提出一种基于多重纹理特征的HEVC帧内编码单元快速划分算法,该算法能够缩小划分的深度区间。首先,使用自定义的纹理提取方法提取出编码单元中的多重纹理特征;其次,使用支持向量机(SVM)对多重纹理特征参数进行训练,得出决策函数;最后,根据决策函数,跳过前面不必要的划分和提前终止划分。实验结果表明,同原始HM 12.0相比,快速划分算法编码时间平均减少43.23%,码率平均增加0.84%,明显提高了帧内编码效率。此外,所提算法容易与其他算法进行融合,进一步降低HEVC的帧内计算复杂度。

关键词: 高效率视频编码, 帧内预测, 编码单元, 四叉树划分, 支持向量机