{"title":"Deep Learning for Biometric Recognition of Children using Footprints","authors":"V. Kamble, M. Dale","doi":"10.1109/ESCI53509.2022.9758315","DOIUrl":null,"url":null,"abstract":"Children are the most important part of society. Every parent is concerned about their health and security. The children's age group of 0 to 5 years is extremely vulnerable. New security options need to be found for the children in this age group. Biometric recognition using their footprint will be an emerging trend for children. This research uses footprint crease pattern of children for recognition. The crease pattern on footprints is extracted for the features. The database of 48 children is collected from preschools and neighborhoods. These images are preprocessed and enhanced. The Transfer learning approach of deep learning is used to compare the proposed method of identification of children. Different deep learning algorithms VGG16, VGG19, ResNet50, AlexNet are used. The proposed method is a fine tuned, customized AlexNet model. The comparison of parameters used is done for all algorithms. Proposed model reduces the number of parameters by 1,69,30,688 with the accuracy of 98 %.","PeriodicalId":436539,"journal":{"name":"2022 International Conference on Emerging Smart Computing and Informatics (ESCI)","volume":"173 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-03-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 International Conference on Emerging Smart Computing and Informatics (ESCI)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ESCI53509.2022.9758315","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
Children are the most important part of society. Every parent is concerned about their health and security. The children's age group of 0 to 5 years is extremely vulnerable. New security options need to be found for the children in this age group. Biometric recognition using their footprint will be an emerging trend for children. This research uses footprint crease pattern of children for recognition. The crease pattern on footprints is extracted for the features. The database of 48 children is collected from preschools and neighborhoods. These images are preprocessed and enhanced. The Transfer learning approach of deep learning is used to compare the proposed method of identification of children. Different deep learning algorithms VGG16, VGG19, ResNet50, AlexNet are used. The proposed method is a fine tuned, customized AlexNet model. The comparison of parameters used is done for all algorithms. Proposed model reduces the number of parameters by 1,69,30,688 with the accuracy of 98 %.