摘要
提出一种改进SIFT算法,该算法主要针对传统SIFT算法数据量大、耗时长的问题,利用主成分分析法对SIFT算法进行改进,降低了SIFT算法提取的特征维数,并结合人脸数据库进行算法验证.结果表明,改进SIFT算法不仅具有亮度变化、旋转和尺度不变性,而且较原算法更稳定、精确、快速.
For a large number of data and long time-consumption of traditional SIFT algorithm, the improved SIFT algorithm is presen ted utilizing principal component analysis method to improve the old SIFT algorithm and reduce the extracted dimension of features. This improved algorithm has been tested and verified combining the face database. The result shows that the improved SIFT algorithn not only has the invariance of brightness variation, rotation and scale, but also is more stable, accurate and rapid than the old algo-rithm.
出处
《华北水利水电学院学报》
2013年第3期125-128,共4页
North China Institute of Water Conservancy and Hydroelectric Power
作者简介
冯博（1985-），男，河南开封人，硕士研究生，主要从事模式识别与图像处理方面的研究．
