论文部分内容阅读
为提高掌纹图像识别率,首先利用手掌的几何轮廓对所采集到的掌纹图像进行预处理,进行分割得到感兴趣的区域。再利用小波变换对掌纹图像分别进行多层分解,进而提取小波特征。最后利用BP神经网络进行分类。通过仿真实验表明,与单一的神经网络方法进行掌纹识别相比,这种将小波分析与神经网络相结合的方法收敛步数少、用时短、具有较高的识别率。
In order to improve the recognition rate of palmprint images, the palm palmprint image is preprocessed by using the geometrical outline of the palms, and the palmprint images are segmented to obtain the regions of interest. The wavelet transform is used to decompose the palmprint images into multi-layers respectively, then the wavelet features are extracted. Finally, using BP neural network to classify. The simulation results show that compared with the single neural network method for palmprint recognition, this method combining wavelet analysis with neural network has fewer convergence steps, shorter time and higher recognition rate.