基于CUDA的PCA-SIFT算法研究
首发时间:2012-12-04
摘要:主成分不变特征检测算法PCA-SIFT被广泛应用于图像特征的检测。本文利用最新的图形处理单元(GPU)并行架构和统一计算设备架构(CUDA)灵活的编程性,提出了一种基于CUDA的快速PCA-SIFT特征检测算法。实验结果表明,与CPU架构下的算法相比,本文设计的算法能够在保证特征检测结果不变的情况下获得3-5倍的加速。
For information in English, please click here
Study of PCA-SIFT algorithm based on CUDA
Abstract:The feature detection algorithms based on the principal component invariant PCA-SIFT are widely used in image feature detection. Using the latest graphics processing unit (GPU) and compute unified device architecture (CUDA) which has flexiblility in the programming, a rapid PCA-SIFT feature detection algorithm based on CUDA is designed in this paper. The experimental results show that compared with the algorithm on CPU architecture, the speedup of the algorithm based on CUDA is 3-5 under the condition of the same results of feature detection.
Keywords: GPU CUDA PCA-SIFT Feature detect
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于CUDA的PCA-SIFT算法研究
评论
全部评论0/1000