| 基于决策树的HEVC编码单元划分算法 |
| |
| 引用本文： | 岑跃峰,王万良,姚信威,王超超,潘铁强. 基于决策树的HEVC编码单元划分算法[J]. 计算机科学, 2016, 43(4): 308-312 |
| |
| 作者姓名： | 岑跃峰 王万良 姚信威 王超超 潘铁强 |
| |
| 作者单位： | 浙江工业大学计算机科学与技术学院 杭州310023,浙江工业大学计算机科学与技术学院 杭州310023,浙江工业大学计算机科学与技术学院 杭州310023,浙江工业大学计算机科学与技术学院 杭州310023,浙江工业大学计算机科学与技术学院 杭州310023;衢州学院电气与信息工程学院 衢州324000 |
| |
| 基金项目： | 本文受国家自然科学基金资助 |
| |
| 摘 要： | 为了有效降低高效视频编码标准HEVC编码过程中的计算复杂度,提出了一种基于决策树的编码单元划分算法。该算法将编码单元划分问题归为分类问题,提取编码单元的划分信息到决策树模型中进行学习,得到决策树分类器。利用分类器对满足分类条件的编码单元进行划分,跳过了率失真计算,从而降低了视频编码的计算复杂度。实验结果表明,本算法在保证视频质量的同时有效降低了编码计算复杂度。
|
| 关 键 词： | HEVC 计算复杂度 决策树 率失真 信噪比 |
| 收稿时间： | 2015-03-19 |
| 修稿时间： | 2015-06-30 |
Decision Tree Based Coding Unit Splitting Algorithm for HEVC |
| |
| CEN Yue-feng,WANG Wan-liang,YAO Xin-wei,WANG Chao-chao and PAN Tie-qiang. Decision Tree Based Coding Unit Splitting Algorithm for HEVC[J]. Computer Science, 2016, 43(4): 308-312 |
| |
| Authors: | CEN Yue-feng WANG Wan-liang YAO Xin-wei WANG Chao-chao PAN Tie-qiang |
| |
| Affiliation: | College of Computer Science & Technology,Zhejiang University of Technology,Hangzhou 310023,China,College of Computer Science & Technology,Zhejiang University of Technology,Hangzhou 310023,China,College of Computer Science & Technology,Zhejiang University of Technology,Hangzhou 310023,China,College of Computer Science & Technology,Zhejiang University of Technology,Hangzhou 310023,China and College of Computer Science & Technology,Zhejiang University of Technology,Hangzhou 310023,China;College of Electrical and Information Engineering,Quzhou University,Quzhou 324000,China |
| |
| Abstract: | To reduce the computational complexity of high efficient video coding (HEVC),a decision tree based coding unit (CU) splitting algorithm was proposed.The splitting of the CU is seen as a classification problem.Furthermore,splitting information extracted from CUs is added to the decision tree model for principle learning.A decision tree based classifier is obtained after the principle learning.Then the classifier is used to determine the splitting of the CU if the classification condition is satisfied.Thus,the computation of rate distortion (RD) cost is skipped,and the encoding computational complexity is reduced.Experimental results demonstrate that the proposed mechanism achieves significant reduction of the encoding computational complexity,while still maintaining high video coding quality. |
| |
| Keywords: | HEVC Computational complexity Decision tree Rate distortion PSNR |
|
|
点击此处可从《计算机科学》下载全文 |
|
