| 基于Chan-Vese模型的面向多核CPU和GPU的人脸轮廓提取并行算法 |
| |
| 引用本文： | 王丽娜,史晓华.基于Chan-Vese模型的面向多核CPU和GPU的人脸轮廓提取并行算法[J].计算机应用,2014,34(11):3121-3125. |
| |
| 作者姓名： | 王丽娜 史晓华 |
| |
| 作者单位： | 北京航空航天大学 计算机学院,北京 100191 |
| |
| 基金项目： | 国家自然科学基金资助项目，国家科技重大专项(核高基) |
| |
| 摘 要： | 针对人脸轮廓提取中Chan-Vese模型计算量大、分割速度缓慢等问题,采用开放计算语言(OpenCL)并行编程模型,提出了一种基于图形处理器(GPU)和多核CPU加速的并行算法。该算法首先将模型的框架进行重构,消除模型中的数据依赖关系;然后,利用开放计算语言对算法进行并行化以及相应的优化。实验结果表明,与单线程算法相比,在NVIDIA GTX660和AMD FX-8530下达到了较高的加速比。
|
| 关 键 词： | Chan-Vese模型 并行 开放计算语言 人脸轮廓提取 |
| 收稿时间： | 2014-07-29 |
| 修稿时间： | 2014-08-07 |
Parallel face contour extraction algorithm based on Chan-Vese model on multi-core CPU and GPU |
| |
| WANG Lina
,
SHI Xiaohua.Parallel face contour extraction algorithm based on Chan-Vese model on multi-core CPU and GPU[J].journal of Computer Applications,2014,34(11):3121-3125. |
| |
| Authors: | WANG Lina SHI Xiaohua |
| |
| Affiliation: | School of Computer Science and Engineering, Beihang University, Beijing 100191, China |
| |
| Abstract: | Aiming at the problem that Chan-Vese model of face contour extraction has a large amount of calculation and slow segmentation, an parallel algorithm was proposed based on Graphic Processing Unit (GPU) and multi-core CPU with Open Computing Language (OpenCL) parallel programming model. Firstly, by reconstructing the computing architecture, the algorithm could eliminate data dependence of the model; secondly, using OpenCL to parallel and optimize the algorithm. The result shows that, compared with the single-thread method, the proposed algorithm gets a higher average acceleration ratios under NVIDIA GTX660 and AMD FX-8530. |
| |
| Keywords: | Chan-Vese model parallel Open Computing Language (OpenCL) face contour extraction |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用》浏览原始摘要信息 |
|
点击此处可从《计算机应用》下载全文 |
|
