生物特征提取的Gabor滤波器参数设计

Design of Gabor filter parameters for biometric feature extraction

为了使Gabor滤波器能够在时域和频域内同时取得局部最优,提出了一种Gabor滤波器的参数设计方法。首先分析了Gabor滤波器方向、尺度参数变化对生物提取特征的影响,然后探讨了滤波器窗口参数变化对提取特征的影响,最后对Gabor滤波器的参数进行优化设计,并将其应用于ORL人脸库的特征提取。结果表明,该方法能使Gabor滤波器在时域和频域均获得最佳滤波性能,有效获得生物特征信息。

In order to obtain local optimality for Gabor filter in bot!i time domain and frequence do-main, a Gabor filters＇ parameter algoritiim was proposed in this paper. ture in Gabor filter diferent directions and diferent scales were analyzed. T h e n, tture in diferent window functions were discussed. Finally, the Gabor parameters was it was used for face biometric feature extraction in ORL face database. The results show rithm can obtain the best filtering performance in the time domain and frequency domain.