Pub Date : 2007-09-11DOI: 10.1109/BCC.2007.4430537
Q. Tao, R. Veldhuis
Illumination normalization is a very important step in face recognition. In this paper we propose a simple implementation of local binary patterns, which effectively reduces the variability caused by illumination changes. In combination with a likelihood ratio classifier, this illumination normalization method achieves very good recognition performance, with respect to both discrimination and generalization. A user verification system using this method has been successfully implemented on a mobile platform.
{"title":"Illumination Normalization Based on Simplified Local Binary Patterns for A Face Verification System","authors":"Q. Tao, R. Veldhuis","doi":"10.1109/BCC.2007.4430537","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430537","url":null,"abstract":"Illumination normalization is a very important step in face recognition. In this paper we propose a simple implementation of local binary patterns, which effectively reduces the variability caused by illumination changes. In combination with a likelihood ratio classifier, this illumination normalization method achieves very good recognition performance, with respect to both discrimination and generalization. A user verification system using this method has been successfully implemented on a mobile platform.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117154067","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430540
Haiping Lu, K. Plataniotis, A. Venetsanopoulos
This paper proposes a novel uncorrelated multilinear discriminant analysis (UMLDA) algorithm for the challenging problem of gait recognition. A tensor-to-vector projection (TVP) of tensor objects is formulated and the UMLDA is developed using TVP to extract uncorrelated discriminative features directly from tensorial data. The small-sample-size (SSS) problem present when discriminant solutions are applied to the problem of gait recognition is discussed and a regularization procedure is introduced to address it. The effectiveness of the proposed regularization is demonstrated in the experiments and the regularized UMLDA algorithm is shown to outperform other multilinear subspace solutions in gait recognition.
{"title":"Uncorrelated Multilinear Discriminant Analysis with Regularization for Gait Recognition","authors":"Haiping Lu, K. Plataniotis, A. Venetsanopoulos","doi":"10.1109/BCC.2007.4430540","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430540","url":null,"abstract":"This paper proposes a novel uncorrelated multilinear discriminant analysis (UMLDA) algorithm for the challenging problem of gait recognition. A tensor-to-vector projection (TVP) of tensor objects is formulated and the UMLDA is developed using TVP to extract uncorrelated discriminative features directly from tensorial data. The small-sample-size (SSS) problem present when discriminant solutions are applied to the problem of gait recognition is discussed and a regularization procedure is introduced to address it. The effectiveness of the proposed regularization is demonstrated in the experiments and the regularized UMLDA algorithm is shown to outperform other multilinear subspace solutions in gait recognition.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126811827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430542
Haiping Lu, K. Plataniotis, A. Venetsanopoulos
In this paper, we present a boosted linear discriminant analysis (LDA) solution with regularization on features extracted by the multilinear principal component analysis (MPCA) for the gait recognition problem. This work is an extension of a recent LDA-based boosting approach and the MPCA is employed to project tensorial gait samples on a number of discriminative EigenTensorGaits (ETGs) to produce gait feature vectors for the base learners in boosting. This new scheme offers one more way to control the learner weakness while being very computationally efficient. Furthermore, the LDA learners are modified through regularization for protection against overfitting on the gallery set. Promising experimental results obtained on the Gait Challenge data sets indicate that the proposed algorithm is an efficient and effective solution consistently enhancing the gait recognition results on the seven probe sets by MPCA+LDA.
{"title":"Boosting LDA with Regularization on MPCA Features for Gait Recognition","authors":"Haiping Lu, K. Plataniotis, A. Venetsanopoulos","doi":"10.1109/BCC.2007.4430542","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430542","url":null,"abstract":"In this paper, we present a boosted linear discriminant analysis (LDA) solution with regularization on features extracted by the multilinear principal component analysis (MPCA) for the gait recognition problem. This work is an extension of a recent LDA-based boosting approach and the MPCA is employed to project tensorial gait samples on a number of discriminative EigenTensorGaits (ETGs) to produce gait feature vectors for the base learners in boosting. This new scheme offers one more way to control the learner weakness while being very computationally efficient. Furthermore, the LDA learners are modified through regularization for protection against overfitting on the gallery set. Promising experimental results obtained on the Gait Challenge data sets indicate that the proposed algorithm is an efficient and effective solution consistently enhancing the gait recognition results on the seven probe sets by MPCA+LDA.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128791104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430550
T. Amin, D. Hatzinakos
This paper presents a new gait signature for human gait recognition which is based on the correlation analysis of the leg motion. The motion of two legs during the human walking process is one of the most important gait determinants. This cyclic motion of the two legs is extracted by applying 2-D masks on the relevant areas of the binary images. Experimental results indicate that 2nd. order and 1-D diagonal slice of 3rd. order autocorrelations of these area signals possesses significant discrimination power to build an effective gait recognition system.
{"title":"A Correlation Based Approach to Human Gait Recognition","authors":"T. Amin, D. Hatzinakos","doi":"10.1109/BCC.2007.4430550","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430550","url":null,"abstract":"This paper presents a new gait signature for human gait recognition which is based on the correlation analysis of the leg motion. The motion of two legs during the human walking process is one of the most important gait determinants. This cyclic motion of the two legs is extracted by applying 2-D masks on the relevant areas of the binary images. Experimental results indicate that 2nd. order and 1-D diagonal slice of 3rd. order autocorrelations of these area signals possesses significant discrimination power to build an effective gait recognition system.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"139 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131976972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430554
Jen-Mei Chang, M. Kirby, C. Peterson
We present a face recognition method using multiple images where pose and illumination are uncontrolled. The set-to-set framework can be utilized whenever multiple images are available for both gallery and probe subjects. We can then transform the set-to-set classification problem as a geometric one by realizing the linear span of the images in a given resolution as a point on the Grassmann manifold where various metrics can be used to quantify the closeness of the identities. Contrary to a common practice, we will not normalize for variations in pose and illumination, hence showing the effectiveness of the set-to-set method when the classification is done on the Grassmann manifold. This algorithm exploits the geometry of the data set such that no training phase is required and may be executed in parallel across large data sets. We present empirical results of this algorithm on the CMU-PIE database and the extended Yale face database B, each consisting of 67 and 28 subjects, respectively.
{"title":"Set-to-Set Face Recognition Under Variations in Pose and Illumination","authors":"Jen-Mei Chang, M. Kirby, C. Peterson","doi":"10.1109/BCC.2007.4430554","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430554","url":null,"abstract":"We present a face recognition method using multiple images where pose and illumination are uncontrolled. The set-to-set framework can be utilized whenever multiple images are available for both gallery and probe subjects. We can then transform the set-to-set classification problem as a geometric one by realizing the linear span of the images in a given resolution as a point on the Grassmann manifold where various metrics can be used to quantify the closeness of the identities. Contrary to a common practice, we will not normalize for variations in pose and illumination, hence showing the effectiveness of the set-to-set method when the classification is done on the Grassmann manifold. This algorithm exploits the geometry of the data set such that no training phase is required and may be executed in parallel across large data sets. We present empirical results of this algorithm on the CMU-PIE database and the extended Yale face database B, each consisting of 67 and 28 subjects, respectively.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134157575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430531
Sung W. Park, M. Savvides
In this paper, we propose a novel method for performing robust super-resolution of face images. Face super-resolution is to recover a high-resolution face image from a given low-resolution face image by modeling a face image space in view of multiple resolutions. The proposed method is based on the assumption that a low-resolution image space and a high-resolution image space have similar local geometries but also have partial distortions of neighborhood relationships between facial images. In this paper, local geometry is analyzed by an idea inspired by locally linear embedding (LLE), the state-of-the art manifold learning method. Using the analyzed neighborhood relationships, two sets of neighborhoods in the low-and high-resolution image spaces become more similar in an iterative way. In this paper, we show that changing resolution causes the partial distortions of neighborhood embeddings obtained by a manifold learning method. Experimental results show that the proposed method produces more reliable results of face super-resolution than the traditional way using neighbor embedding.
{"title":"Robust Super-Resolution of Face Images by Iterative Compensating Neighborhood Relationships","authors":"Sung W. Park, M. Savvides","doi":"10.1109/BCC.2007.4430531","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430531","url":null,"abstract":"In this paper, we propose a novel method for performing robust super-resolution of face images. Face super-resolution is to recover a high-resolution face image from a given low-resolution face image by modeling a face image space in view of multiple resolutions. The proposed method is based on the assumption that a low-resolution image space and a high-resolution image space have similar local geometries but also have partial distortions of neighborhood relationships between facial images. In this paper, local geometry is analyzed by an idea inspired by locally linear embedding (LLE), the state-of-the art manifold learning method. Using the analyzed neighborhood relationships, two sets of neighborhoods in the low-and high-resolution image spaces become more similar in an iterative way. In this paper, we show that changing resolution causes the partial distortions of neighborhood embeddings obtained by a manifold learning method. Experimental results show that the proposed method produces more reliable results of face super-resolution than the traditional way using neighbor embedding.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131773404","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430536
C. Busch, A. Nouak, Xuebing Zhou, F. Deravi, M. van der Veen, Jean-Marc Suchier
Biometric data have been integrated in all new European passports, since the member states of the European Union started to implement the EU Council Regulation No 2252/2004 on standards for security features and biometrics in passports. The additional integration of three-dimensional facial models promises significant performance enhancements for border control applications. By combining the geometry-and texture-channel information of the face, 3D face recognition systems provide improved robustness while being able to handle variations in poses and problematic lighting conditions during image acquisition. To assess the potential of three-dimensional face recognition, the 3D Face Integrated Project was initiated as part of the European Framework Program for collaborative research in April 2006. This paper outlines the research objectives and the approach of this project: Not only shall the recognition performance be increased but also a new, fake-resistant acquisition system is to be developed. In addition, methods for protection of the stored template data in the biometric reference are under development to enhance the privacy and security of the overall system. The use of multi-biometrics is also a key feature of the 3D Face project addressing the performance, robustness and flexibility targets of the system.
{"title":"Towards Unattended and Privacy Protected Border Control","authors":"C. Busch, A. Nouak, Xuebing Zhou, F. Deravi, M. van der Veen, Jean-Marc Suchier","doi":"10.1109/BCC.2007.4430536","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430536","url":null,"abstract":"Biometric data have been integrated in all new European passports, since the member states of the European Union started to implement the EU Council Regulation No 2252/2004 on standards for security features and biometrics in passports. The additional integration of three-dimensional facial models promises significant performance enhancements for border control applications. By combining the geometry-and texture-channel information of the face, 3D face recognition systems provide improved robustness while being able to handle variations in poses and problematic lighting conditions during image acquisition. To assess the potential of three-dimensional face recognition, the 3D Face Integrated Project was initiated as part of the European Framework Program for collaborative research in April 2006. This paper outlines the research objectives and the approach of this project: Not only shall the recognition performance be increased but also a new, fake-resistant acquisition system is to be developed. In addition, methods for protection of the stored template data in the biometric reference are under development to enhance the privacy and security of the overall system. The use of multi-biometrics is also a key feature of the 3D Face project addressing the performance, robustness and flexibility targets of the system.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116629765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430556
F. Bui, D. Hatzinakos
Body sensor networks (BSN) have the potential to provide improved data collection and analysis, as well as enhanced security, particularly in a wide range of medical applications. One of the main challenges in these types of networks is scarce resources, in terms of both computational and communication capabilities. In this work, we present methods to efficiently allocate these limited resources, while maintaining good security performance. Two main strategies are explored: first, a key distribution system is presented that allows for trade-offs between computational complexity and spectral efficiency; second, a data scrambling method based on random sampling is proposed as a possible alternative to conventional encryption in providing security. The obtained simulation results demonstrate the feasibility and efficacy of these schemes in the context of BSN, when using electrocardiogram (ECG) signals as biometrics.
{"title":"Resource Allocation Strategies for Secure and Efficient Communications in Biometrics-Based Body Sensor Networks","authors":"F. Bui, D. Hatzinakos","doi":"10.1109/BCC.2007.4430556","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430556","url":null,"abstract":"Body sensor networks (BSN) have the potential to provide improved data collection and analysis, as well as enhanced security, particularly in a wide range of medical applications. One of the main challenges in these types of networks is scarce resources, in terms of both computational and communication capabilities. In this work, we present methods to efficiently allocate these limited resources, while maintaining good security performance. Two main strategies are explored: first, a key distribution system is presented that allows for trade-offs between computational complexity and spectral efficiency; second, a data scrambling method based on random sampling is proposed as a possible alternative to conventional encryption in providing security. The obtained simulation results demonstrate the feasibility and efficacy of these schemes in the context of BSN, when using electrocardiogram (ECG) signals as biometrics.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134313523","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430535
G. Becker, M. Potts
The goal of this research is to demonstrate how a non-metric clustering technique can be used to effectively reduce the search time for finding matches among biometric templates. Some biometric modalities (such as fingerprint) have proven to not cluster effectively with traditional clustering techniques. Without clustering, identification requires an expensive exhaustive search. This research explores the effectiveness of a novel clustering technique using false matches in a non-metric space. False matches are typically undesirable false positive errors that increase with gallery size. This clustering approach uses these false matches as references for clustering in non-metric similarity space. Searches can then be restricted to only those clusters that claim the probe as a member.
{"title":"Non-Metric Biometric Clustering","authors":"G. Becker, M. Potts","doi":"10.1109/BCC.2007.4430535","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430535","url":null,"abstract":"The goal of this research is to demonstrate how a non-metric clustering technique can be used to effectively reduce the search time for finding matches among biometric templates. Some biometric modalities (such as fingerprint) have proven to not cluster effectively with traditional clustering techniques. Without clustering, identification requires an expensive exhaustive search. This research explores the effectiveness of a novel clustering technique using false matches in a non-metric space. False matches are typically undesirable false positive errors that increase with gallery size. This clustering approach uses these false matches as references for clustering in non-metric similarity space. Searches can then be restricted to only those clusters that claim the probe as a member.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114653781","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2007-09-01DOI: 10.1109/BCC.2007.4430543
E. Maiorana, P. Campisi, A. Neri
In this paper we propose a signature-based biometric authentication system, where watermarking techniques are used to embed some dynamic signature features in a static representation of the signature itself. User authentication can be performed either by means of the only signature static image, or by using it together with the dynamic features embedded in the enrollment stage, by using a fusion approach. A multilevel authentication system, which is capable to provide two different levels of security, is then obtained. The proposed watermarking techniques are based on the properties of the Radon transform which well fits to the signature images. A robust embedding is obtained while keeping unaltered the original structure of the host signal.
{"title":"Biometric Signature Authentication Using Radon Transform-Based Watermarking Techniques","authors":"E. Maiorana, P. Campisi, A. Neri","doi":"10.1109/BCC.2007.4430543","DOIUrl":"https://doi.org/10.1109/BCC.2007.4430543","url":null,"abstract":"In this paper we propose a signature-based biometric authentication system, where watermarking techniques are used to embed some dynamic signature features in a static representation of the signature itself. User authentication can be performed either by means of the only signature static image, or by using it together with the dynamic features embedded in the enrollment stage, by using a fusion approach. A multilevel authentication system, which is capable to provide two different levels of security, is then obtained. The proposed watermarking techniques are based on the properties of the Radon transform which well fits to the signature images. A robust embedding is obtained while keeping unaltered the original structure of the host signal.","PeriodicalId":389417,"journal":{"name":"2007 Biometrics Symposium","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127729993","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：