节点文献
虹膜识别中的一种神经算法的研究
A NEURAL ALGORITHM APPROACH TO IRIS RECOGNITON
【摘要】 研究了与传统身份识别不同的虹膜识别 ,在获取虹膜采样图后 ,经小波变换、松弛神经网络算法及稀疏编码 ,形成虹膜纹理的代码。并以海明距离作为决策空间的判决标准 ,实现人的虹膜图像识别的全过程 ,并对一些特定干扰条件下的虹膜图像进行计算机模拟试验 ,取得较好的结果。
【Abstract】 Extracted from human iris samples, with the mathematical model of concentric receptive field as the mother function, iris texture are encoded sparsely by wavelet transform whose coefficients are obtained by relaxed neural network. And Hamming Distance between the input iris and the iris database is introduced in decision making. We simulated the whole process of iris recognition on SGI workstation and get reliable results, especially in those images with some certain disturbance.
【关键词】 感受野;
虹膜识别;
虹膜代码;
小波变换;
松弛网络;
稀疏编码;
零交叉;
【Key words】 Receptive field; Iris Recognition; Iriscode; Wavelet transform; Relaxation network; Sparse coding; Zero crossings|;
【Key words】 Receptive field; Iris Recognition; Iriscode; Wavelet transform; Relaxation network; Sparse coding; Zero crossings|;
【基金】 国家自然科学基金委项目资助!项目号为39893340 -06 ;3967086;69835020
- 【文献出处】 生物物理学报 ,Acta Biophysica Sinica , 编辑部邮箱 ,2000年04期
- 【分类号】Q811.3
- 【被引频次】51
- 【下载频次】207