摘要
运用传统方法对模糊指纹图像奇异点进行检测时存在误差率较高和漏检率较大等问题,为此提出了基于Contourlet变换和模糊逻辑的模糊指纹图像奇异点检测方法。运用相对梯度和绝对梯度相融合的方法,增强模糊指纹图像较亮区域的梯度,利用矩阵乘法与求逆算法进行离散正弦变换,构建人工智能辅助下模糊指纹图像增强模型,并对该模型进行Contourlet转换,获取模糊指纹图像信号尺度和方向上的低频和高频变换系数,将该变换系数当做语言变量输入,利用模糊逻辑方法计算各个模糊区域所激活的强度值,将其归一化检测后,输出模糊指纹图像奇异点。分析实验结果可知,所提方法的最低漏检率为2%,远低于传统方法,说明该方法能够增强检测的准确率、降低漏检率和误差率,具备一定的可靠性。
Traditionally,there are some problems such as high error rate and high missing detection rate when detecting the singular point of fuzzy fingerprint image.Therefore,a method to detect the singular point of fuzzy fingerprint image based on Contourlet transform and fuzzy logic was proposed.Combined relative gradient with absolute gradient,the gradient of brighter region in fuzzy fingerprint image was enhanced.Then,matrix multiplication and inversion algorithm were applied to the discrete sine transform.After that,the enhancement model of fuzzy fingerprint image based on artificial intelligent assistant was built.Meanwhile,Contourlet transformation was performed on this model,so as to obtain the low frequency and high frequency transformation coefficients on the direction and signal scale of fuzzy fingerprint image,which were regarded as the language variable input.Moreover,the fuzzy logic method was used to calculate the intensity value activated by each fuzzy region.After normalization detection,the singular points of fuzzy fingerprint image were output.Simulation results show that the minimum miss detection rate of the proposed method is 2%,which is much lower than that of traditional methods.Therefore,the proposed method can enhance the detection accuracy rate and reduce the miss detection rate and error rate,which has certain reliability.
作者
刘强
周波
LIU Qiang;ZHOU Bo(Hefei University Of Technology School of Computer and Information,AnHui,Xuancheng,242000,China)
出处
《计算机仿真》
北大核心
2020年第2期426-429,485,共5页
Computer Simulation
关键词
指纹图像
奇异点检测
位置误差率
降低漏检率
可靠性
Fingerprint image
Singular point detection
Position error rate
Reduce miss detection rate
Reliability