关于人脸图像抗噪性能检测算法研究  

Antinoise Performance of Face Image Detection Algorithm Research

作　　者：胡智鹏[1] 刘颖[1] 

机构地区：[1]北华大学信息技术与传媒学院,吉林吉林132013

出　　处：《计算机仿真》2016年第11期349-352,共4页Computer Simulation

摘　　要：对人脸图像抗噪性能进行检测,可以为图像后续处理提高检测精确度。在抗噪性能检测过程中通过将人脸图像中噪声点的分布情况进行分析,然后利用邻域加权中值对图像中的噪声点进行降维检测而实现的,而传统的蚁群算法只包含简单的隶属度函数,不能计算邻域加权中值,降低了人脸图像抗噪性能的检测精度和效率。提出一种多目标优化算法的人脸图像抗噪性能检测算法。先利用模糊推理假设规则将待处理图像映射为一个模糊矩阵,利用矩阵分割中的阈值选取方法计算阈值参数,根据计算结果将人脸图像的直方图分为高、低灰度区域,并对高灰度图像进行模糊增强处理,利用模糊边缘检测算法获取人脸图像边缘信息,并设计图像检测的自适应模糊滤波器,在不同调整图像边缘检测策略的条件下对噪声强度进行计算,实现人脸图像抗噪性能检测。仿真结果表明,改进的人脸图像抗噪性能检测算法提高了检测精度和效率。An anti-noise detection method of facial image based on multi-objective optimization algorithm is pro- posed. Firstly, the rule of fuzzy reasoning hypothesis is used to map the pending image into a fuzzy matrix, and the threshold selection method in matrix partition is used to calculate the threshold parameter. According to the calculation results, the histogram of facial image is divided into high and low gray areas, and the high gray image is processed by fuzzy enhancement. The fuzzy edge detection algorithm is used to obtain the edge information of the facial image, the adaptive fuzzy filter of image detection is designed, the noise intensity is calculated, and anti-noise performance detection of facial image is achieved under the condition of different image edge detection strategies. The simulation results show that the improved algorithm can increase the detection accuracy and efficiency.

关 键 词：图像模糊 边缘检测 噪声 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]