基于加权强度PCNN模型的分块人脸识别
|
摘要
为提高利用脉冲耦合神经网络(pulse coupled neural network,PCNN)进行人脸识别时的准确率,提出基于加权强度PCNN模型的分块人脸识别方法。在简化PCNN模型的基础上,引入自发脉冲发放强度、耦合脉冲发放强度和加权强度的概念,细化模型的输出;根据人脸图像各部分灰度分布的不同和局部识别率的不同,将人脸图像进行分块;进行人脸识别时,分块的权值会根据分块图像的局部信息熵自适应地设定,模型的参数会根据分块图像内容设定,一幅人脸图像的识别结果会综合各分块的识别结果。多个数据库上的实验结果表明,该算法可以有效地提高识别率。
To raise the accuracy of face recognition using pulse coupled neural network(PCNN)model,the block face recognition method using PCNN model based on weighted strength was proposed.Based on the simplified PCNN model,spontaneous pulse strength,coupled pulse strength and weighted strength were proposed,which made the output of the model more accurate.Each face image was divided into several block images according to the recognition rate and gray distribution of different block images.The weight corresponding to one block image was set self-adaptively according to its local entropy.Parameters of PCNN model were also set according to different block images.All recognition results corresponding to all block images of one face image were comprehensively considered to get the final recognition result.Experimental results show that the proposed method can improve face recognition rate effectively.
To raise the accuracy of face recognition using pulse coupled neural network(PCNN)model,the block face recognition method using PCNN model based on weighted strength was proposed.Based on the simplified PCNN model,spontaneous pulse strength,coupled pulse strength and weighted strength were proposed,which made the output of the model more accurate.Each face image was divided into several block images according to the recognition rate and gray distribution of different block images.The weight corresponding to one block image was set self-adaptively according to its local entropy.Parameters of PCNN model were also set according to different block images.All recognition results corresponding to all block images of one face image were comprehensively considered to get the final recognition result.Experimental results show that the proposed method can improve face recognition rate effectively.
引文
[1]KOU Guangjie,MA Yunyan,YUE Jun,et al.Survey of bioinspired natural computing[J].Computer Science,2014,41(6A):37-41(in Chinese).[寇光杰,马云艳,岳峻,等.仿生自然计算研究综述[J].计算机科学,2014,41(6A):37-41.]
[2]SHAO Xiaopeng,ZHONG Cheng,WANG Yang,et al.Application of a simplified PCNN model in color image edge detection[J].Journal of Xidian University,2012,39(6):1-9(in Chinese).[邵晓鹏,钟宬,王杨,等.一种简化PCNN模型在彩色图像边缘检测上的应用[J].西安电子科技大学学报,2012,39(6):1-9.]
[3]WANG Li,WANG Min.Automatic image segmentation algorithm by PCNN based on mean threshold and ostu[J].Infrared Technology,2015,37(7):553-559(in Chinese).[王力,王敏.基于区间参数寻优的PCNN红外图像自动分割方法[J].红外技术,2015,37(7):553-559.]
[4]Jiang Wen,Zhou Haoyu,Shen Yue,et al.Image segmentation with pulse-coupled neural network and canny operators[J].Computers and Electrical Engineering,2015,46:528-538.
[5]JIANG Zhi.Image fusion algorithm based on PCNN and NSCT transform[D].Hangzhou:Zhejiang Sci-Tech University,2014:26-43(in Chinese).[姜志.基于PCNN和NSCT变换的图像融合算法[D].杭州:浙江理工大学,2014:26-43.]
[6]LI Xi,ZHENG Hong,LIU Cao.A method of facerecognition based on HIS-PCNN[J].Geomatics and Information Science of Wuhan University,2014,39(12):1499-1503(in Chinese).[黎曦,郑宏,刘操.利用HSI色彩空间的PCNN人脸识别方法[J].武汉大学学报:信息科学版,2014,39(12):1499-1503.]
[7]YUAN Gang.Application research of PCNN in image denoising and face recognition[D].Kunming:Yunnan University,2015:24-38(in Chinese).[袁刚.脉冲耦合神经网络在图像去噪和人脸识别中的应用研究[D].昆明:云南大学,2015:24-38.]
[8]YANG Xue,LIU Tianshi,LI Xiangjuan.Study on image enhancement algorithm merged wavelet transform and improved PCNN[J].Computer Engineering and Applications,2016,52(8):163-168(in Chinese).[杨雪,刘天时,李湘眷.融合小波变换与改进PCNN的图像增强算法研究[J].计算机工程与应用,2016,52(8):163-168.]
[9]Jin Xin,Nie Rencan,Zhou Dongming,et al.A novel DNA sequence similarity calculation based on simplified pulse-coupled neural network and Huffman coding[J].Physica A Statistical Mechanics&Its Applications,2016,461:325-338.
[10]WANG Xuan,YANG Guang.Modified method for facial feature extraction based on pulse-coupled neural network[J].Computer Engineering and Applications,2013,49(1):213-216(in Chinese).[王晅,杨光.改进的脉冲耦合神经网络人脸特征提取方法[J].计算机工程与应用,2013,49(1):213-216.]
[2]SHAO Xiaopeng,ZHONG Cheng,WANG Yang,et al.Application of a simplified PCNN model in color image edge detection[J].Journal of Xidian University,2012,39(6):1-9(in Chinese).[邵晓鹏,钟宬,王杨,等.一种简化PCNN模型在彩色图像边缘检测上的应用[J].西安电子科技大学学报,2012,39(6):1-9.]
[3]WANG Li,WANG Min.Automatic image segmentation algorithm by PCNN based on mean threshold and ostu[J].Infrared Technology,2015,37(7):553-559(in Chinese).[王力,王敏.基于区间参数寻优的PCNN红外图像自动分割方法[J].红外技术,2015,37(7):553-559.]
[4]Jiang Wen,Zhou Haoyu,Shen Yue,et al.Image segmentation with pulse-coupled neural network and canny operators[J].Computers and Electrical Engineering,2015,46:528-538.
[5]JIANG Zhi.Image fusion algorithm based on PCNN and NSCT transform[D].Hangzhou:Zhejiang Sci-Tech University,2014:26-43(in Chinese).[姜志.基于PCNN和NSCT变换的图像融合算法[D].杭州:浙江理工大学,2014:26-43.]
[6]LI Xi,ZHENG Hong,LIU Cao.A method of facerecognition based on HIS-PCNN[J].Geomatics and Information Science of Wuhan University,2014,39(12):1499-1503(in Chinese).[黎曦,郑宏,刘操.利用HSI色彩空间的PCNN人脸识别方法[J].武汉大学学报:信息科学版,2014,39(12):1499-1503.]
[7]YUAN Gang.Application research of PCNN in image denoising and face recognition[D].Kunming:Yunnan University,2015:24-38(in Chinese).[袁刚.脉冲耦合神经网络在图像去噪和人脸识别中的应用研究[D].昆明:云南大学,2015:24-38.]
[8]YANG Xue,LIU Tianshi,LI Xiangjuan.Study on image enhancement algorithm merged wavelet transform and improved PCNN[J].Computer Engineering and Applications,2016,52(8):163-168(in Chinese).[杨雪,刘天时,李湘眷.融合小波变换与改进PCNN的图像增强算法研究[J].计算机工程与应用,2016,52(8):163-168.]
[9]Jin Xin,Nie Rencan,Zhou Dongming,et al.A novel DNA sequence similarity calculation based on simplified pulse-coupled neural network and Huffman coding[J].Physica A Statistical Mechanics&Its Applications,2016,461:325-338.
[10]WANG Xuan,YANG Guang.Modified method for facial feature extraction based on pulse-coupled neural network[J].Computer Engineering and Applications,2013,49(1):213-216(in Chinese).[王晅,杨光.改进的脉冲耦合神经网络人脸特征提取方法[J].计算机工程与应用,2013,49(1):213-216.]