摘要
通常的指纹识别算法因为预处理步骤过于复杂而明显地存在计算量过大、识别速度慢等缺点。提出一种基于小波变换和对传网络的指纹识别算法,直接从二值化指纹图像中提取指纹图象细节特征进行比对,避免了复杂的预处理过程,减少了工作量,加快了计算速度。同时对对传神经网络进行改进,在竞争层使用软竞争替代传统的硬竞争,以便于分类识别,有效提高识别精度。实验表明该算法在实际应用中效果较好。对受噪声污染严重的指纹图像亦能获得很好的识别效果,算法具有较强的鲁棒性。
In the fingerprint identification system, the pre -processing usually includes image enhancement, smoothing,binarization and thinning. In this paper a new method of fingerprint recognition with no regquirement on complicated pre- processing is proposed , which is based on wavelet transform and CPN （CounterPropagation Neural Network）. The learning mechanism of soft - competition is used in the competition layer replacing the winner - take - all mechanism of tradititonal CPN to get better performance. Experiments show that the proposed method is efective, and more robust than other algorithm.
出处
《计算机仿真》
CSCD
2006年第11期198-201,共4页
Computer Simulation
关键词
指纹识别
小波变换
对传网络
Fingerprint identification
Wavelet transform
CPN
