摘要
提出了一种基于互功率谱的虹膜编码方法。采用粗定位与精定位相结合的两步定位法对虹膜定位,能有效地减少搜索计算的盲目性;为了提取虹膜特征,虹膜图像首先被划分为许多子块,然后计算各个子块与2维Gabor小波族之间的互功率谱并进行编码;最后,利用Hamming距离进行模式匹配。实验结果表明,该算法运算速度快,具有较好的识别效果,且编码性能优于Daugman的编码方法。
An encoding algorithm based on cross power spectrum is proposed. A two-step locating approach combining coarse location with fine location is used to locate the iris, which can effectively decrease the blindness of searching computation. To extract the feature of the iris, the iris image is firstly divided into many blocks, and then the cross power spectra between each block and 2D Gabor wavelet family are computed and the blocks are encoded. Finally, the pattern match is executed by using Hamming distance. The results show that this approach has very high computational speed and good recognition rate and is superior to Daugman＇s method in coding performance.
出处
《中国图象图形学报》
CSCD
北大核心
2007年第5期854-859,共6页
Journal of Image and Graphics
关键词
虹膜识别
虹膜定位
互功率谱
GABOR小波
iris recognition, iris location, cross power spectrum, Gabor wavelets
