A GPU-based Affine and Scale Invariant Feature Transform Algorithm
College of Information Science and Engineering, Northeastern University, Shenyang 110004, China
Journal of Information & Computational Science 10:13, 3991-3998, 2013
@article{tong2013gpu,
title={A GPU-based Affine and Scale Invariant Feature Transform Algorithm},
author={Tong, Guofeng and Meng, Xiangnan and Ye, Ning},
year={2013}
}
Affine invariance is one of the main performances of a good feature extraction algorithm. SIFT is a kind of scale-invariant feature extraction algorithm, but it is not affine invariant. In order to improve SIFT algorithm’s affine invariance. Affine and Scale Invariant Feature Transform (ASIFT) algorithm takes affine Model into SIFT. However, serial ASIFT algorithm’s computing rate is low and it is not suitable for interactive applications. By analyzing the algorithm’s principle of ASIFT, this paper proposed multicore parallel ASIFT algorithm and achieved ASIFT algorithm’s parallelization by CUDA architecture. Finally, the experiment shown that multicore parallel ASIFT algorithm greatly improves computing speed in the same precision as mononuclear serial ASIFT algorithm.
September 14, 2013 by hgpu
