摘要
针对利用单一方法进行掌纹图像识别所得的识别率难以提高这一情况,提出一种利用掌纹图像经高斯高通滤波后的局部二进制模式特征和三级小波分解的细节图像的能量特征的融合特征进行掌纹识别的方法。在提取图像的局部二进制模式特征的时候,通过高斯高通滤波增强图像的对比度,从而提取出更有效的局部二进制模式特征,该特征对光照的变化具有一定的鲁棒性;小波变换的细节图像能量数据反映不同频率成分的局部细节特征。实验结果表明所提出的掌纹识别方法的有效性。
For the situation of using a single method to recognized palmprint images always difficult to improve the rec- ognition rate, a palmprint recognition method which uses the fused feature was proposed. This method combines the Local Binary Pattern（LBP） feature from the image after gaussian high--pass filtering with the energy data from the detail images through three--level wavelet decomposition. When extracting the LBP feature of image, gaussian high--pass filtering can enhance the contrastion of image, and it brings a more effective LBP feature; this feature has some robustness on the chan- ging of illumination; the energy data of detail images from wavelet transform reflect the local detail features of the images under different frequncy component. The experimental results show the effectiveness of the proposed palrnprint recognition method.
出处
《计算技术与自动化》
2010年第1期102-105,共4页
Computing Technology and Automation
