摘要
判断一种算法的GPU实现是否和CPU实现一样可靠、算法的GPU实现是否能够被安全地使用,是GPU实现能否进一步推广的关键问题.文中以SIFT和KD树搜索这2个被广泛使用的算法为例,对其GPU实现和CPU实现进行了系统的实验分析和比较测试.此外,针对高分辨率图像,给出了基于GPU的SIFT算法和KD树搜索算法的设计与实现.实验结果表明,基于GPU的实现和CPU的实现能提取出95%左右的相似特征点;2种算法实现的特征点重复率和匹配率相当,但GPU实现可获得高达13X(特征提取)和43X(特征匹配)的加速比,因此基于GPU的实现在具体应用中完全可以被安全地使用.
Due to the various limits in the GPU implementation, questions may arise. Is the GPU implementation as faithful as the CPU one.＊ Could the GPU implementation be safely used？ In this paper, we take two widely used algorithms as representative examples to evaluate the performance of the GPU implementation, that is, the scale-invariant feature transform （SIFT） for feature detection and the KD-tree traversal for feature matching. We present the design, implementation, and evaluation of these two algorithms on GPU for high-resolution image datasets. Our results show that around 95~ of the extracted {eatures are nearly the same, and the repeatability score and matching score are similar under the GPU and CPU implementations under various image changes. The run-time speedup of the GPU implementation for the SIFT detection is about 13X faster than their CPU counterpart, and that for the KD-tree traversal is about 43X. In sum, our results show that the GPU implementations for feature detection and matching are as good as the CPU ones, and can be safely used in real applications.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2013年第10期1496-1502,共7页
Journal of Computer-Aided Design & Computer Graphics
基金
中国科学院战略性先导科技专项(XDA06030300)
福建省科技厅重点项目(2011H0032)
