Several mathematical color models have been proposed to segment images based on their color information content. The most frequently used color models of such sort include RGB, HSV, YCbCr, etc. These models were designed to represent color and in some cases emulate how the reflection of light on a given entity is perceived by the human eye. They were, however, not designed specifically for the purpose of image segmentation. In this study, the efficiency of several color models for the application of image segmentation is assessed and more efficient color models, consisting of color model mixtures, are explored. It was observed that two of the studied models, YCbCr and linear, were more efficient for the purpose of image segmentation. Additionally, by employing multivariate analysis, it was observed that the model mixtures were more efficient than the most commonly used models studied, and thus optimized the segmentation
{"title":"Optimizing image segmentation using color model mixtures","authors":"Aristide C. Chikando, J. Kinser","doi":"10.1109/AIPR.2005.38","DOIUrl":"https://doi.org/10.1109/AIPR.2005.38","url":null,"abstract":"Several mathematical color models have been proposed to segment images based on their color information content. The most frequently used color models of such sort include RGB, HSV, YCbCr, etc. These models were designed to represent color and in some cases emulate how the reflection of light on a given entity is perceived by the human eye. They were, however, not designed specifically for the purpose of image segmentation. In this study, the efficiency of several color models for the application of image segmentation is assessed and more efficient color models, consisting of color model mixtures, are explored. It was observed that two of the studied models, YCbCr and linear, were more efficient for the purpose of image segmentation. Additionally, by employing multivariate analysis, it was observed that the model mixtures were more efficient than the most commonly used models studied, and thus optimized the segmentation","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"189 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123514724","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}
Activity detection (e.g. recognizing people's behavior and intent), when used over an extended range of applications, suffers from high false detection rates. Also, activity detection limited to 2D image domain (symbolic space) is confined to qualitative activities. Symbolic features, represented by apparent dimensions, i.e. pixels, can vary with distance or viewing angle. One way to enhance performance is to work within the physical space, where object features are represented by their physical dimensions (e.g. inches or centimeters) and are invariant to distance or viewing angle. In this paper, we propose an approach to construct a 3D site model and co-register the video with the site model to obtain real-time physical reference at every pixel in the video. We present a unique approach that creates a 3D site model via fusion of laser range sensor and a single camera. We present experimental results to demonstrate our approach.
{"title":"3D scene modeling using sensor fusion with laser range finder and image sensor","authors":"Yunqian Ma, Z. Wang, Michael E. Bazakos, W. Au","doi":"10.1109/AIPR.2005.5","DOIUrl":"https://doi.org/10.1109/AIPR.2005.5","url":null,"abstract":"Activity detection (e.g. recognizing people's behavior and intent), when used over an extended range of applications, suffers from high false detection rates. Also, activity detection limited to 2D image domain (symbolic space) is confined to qualitative activities. Symbolic features, represented by apparent dimensions, i.e. pixels, can vary with distance or viewing angle. One way to enhance performance is to work within the physical space, where object features are represented by their physical dimensions (e.g. inches or centimeters) and are invariant to distance or viewing angle. In this paper, we propose an approach to construct a 3D site model and co-register the video with the site model to obtain real-time physical reference at every pixel in the video. We present a unique approach that creates a 3D site model via fusion of laser range sensor and a single camera. We present experimental results to demonstrate our approach.","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"105 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116117812","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}
Transfer of digital medical images between multiple parties requires the assurance of image identity and integrity, which can be achieved through image watermarking. This raises concerns for loss in viewer performance due to degradation of image quality. Here we describe an approach to ensure that impact on the image quality is well below the threshold of visual perceptibility. The principles on which this approach rests are the choice of a suitably light payload, and the use of different watermarking methods and parameters for different medical image types. We provide examples of this approach applied to MR, CT and CR images
{"title":"Medical image watermarking for multiple modalities","authors":"A. Maeder, B. Planitz","doi":"10.1109/AIPR.2005.33","DOIUrl":"https://doi.org/10.1109/AIPR.2005.33","url":null,"abstract":"Transfer of digital medical images between multiple parties requires the assurance of image identity and integrity, which can be achieved through image watermarking. This raises concerns for loss in viewer performance due to degradation of image quality. Here we describe an approach to ensure that impact on the image quality is well below the threshold of visual perceptibility. The principles on which this approach rests are the choice of a suitably light payload, and the use of different watermarking methods and parameters for different medical image types. We provide examples of this approach applied to MR, CT and CR images","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123709186","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}
Biometric systems based solely on one-modal biometrics are often not able to meet the desired performance requirements for large user population applications, due to problems such as noisy data, intra-class variations, restricted degrees of freedom, nonuniversity, spoof attacks, and unacceptable error rates. Multimodal biometrics refers to the use of a combination of two or more biometric modalities in a single identification system. The most compelling reason to combine different modalities is to improve the recognition accuracy. This can be done when features of different biometrics are statistically independent. This paper overviews and discusses the various scenarios that are possible in multimodal biometric systems using fingerprint, face and iris recognition, the levels of fusion that are possible and the integration strategies that can be adopted to fuse information and improve overall system accuracy. This paper also discusses how the image quality of fingerprint, face and iris used in the multimodal biometric systems affects the overall identification accuracy and the need of staffing for the secondary human validation. For a large user population identification system, which often has more than tens or hundreds of millions of subject images already enrolled in the matcher databases and has to process more than hundreds of thousands of identification requests, the system's identification accuracy and the need of staffing levels to properly operate the system are two of the most important factors in determining whether a system is properly designed and integrated
{"title":"Multimodal biometric identification for large user population using fingerprint, face and iris recognition","authors":"Teddy Ko","doi":"10.1109/AIPR.2005.35","DOIUrl":"https://doi.org/10.1109/AIPR.2005.35","url":null,"abstract":"Biometric systems based solely on one-modal biometrics are often not able to meet the desired performance requirements for large user population applications, due to problems such as noisy data, intra-class variations, restricted degrees of freedom, nonuniversity, spoof attacks, and unacceptable error rates. Multimodal biometrics refers to the use of a combination of two or more biometric modalities in a single identification system. The most compelling reason to combine different modalities is to improve the recognition accuracy. This can be done when features of different biometrics are statistically independent. This paper overviews and discusses the various scenarios that are possible in multimodal biometric systems using fingerprint, face and iris recognition, the levels of fusion that are possible and the integration strategies that can be adopted to fuse information and improve overall system accuracy. This paper also discusses how the image quality of fingerprint, face and iris used in the multimodal biometric systems affects the overall identification accuracy and the need of staffing for the secondary human validation. For a large user population identification system, which often has more than tens or hundreds of millions of subject images already enrolled in the matcher databases and has to process more than hundreds of thousands of identification requests, the system's identification accuracy and the need of staffing levels to properly operate the system are two of the most important factors in determining whether a system is properly designed and integrated","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121855627","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}
We propose a method that not only identifies humans in the environment and their location, but can also classify and identify their activity, providing a threat assessment. Such assessments would be useful for both human and vehicle activities in crowds to determine aberrant behavior from previously identified truth data sets. Such aberrant behavior would lead to IED detection, RPG detection, and the recognition of suicide bombers, before the explosives and planted and activated. The heuristics needed involve recognition of information bearing features in the environment, and the determination of how those features relate to each other over time (that is, gesture recognition). This paper addresses the mathematical development necessary to create a behavior and gait recognition sensor system that has its foundation on the recognition of combined individual gestures
{"title":"A control theoretic method for categorizing visual imagery as human motion behaviors","authors":"C. Cohen","doi":"10.1109/AIPR.2005.6","DOIUrl":"https://doi.org/10.1109/AIPR.2005.6","url":null,"abstract":"We propose a method that not only identifies humans in the environment and their location, but can also classify and identify their activity, providing a threat assessment. Such assessments would be useful for both human and vehicle activities in crowds to determine aberrant behavior from previously identified truth data sets. Such aberrant behavior would lead to IED detection, RPG detection, and the recognition of suicide bombers, before the explosives and planted and activated. The heuristics needed involve recognition of information bearing features in the environment, and the determination of how those features relate to each other over time (that is, gesture recognition). This paper addresses the mathematical development necessary to create a behavior and gait recognition sensor system that has its foundation on the recognition of combined individual gestures","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126583576","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}
M. Mahfouz, A. Fathy, Yunqiang Yang, Emam ElHak Ali, A. Badawi
Surveillance/navigation systems presently used make extensive use of television, infrared, and other line-of-sight-surveillance hardware. However, these systems cannot tell what is happening or locate persons/assets on the other side of a wall, behind bushes, in the dark, in a tunnel or a cave, or through a dense fog. It is our objective here to develop a new sensor, based on UWB technology. A small, lightweight, low power transceiver or multiples that are based upon the fact that microwave frequencies can be optimized to penetrate nonmetallic materials, and providing very precise ranging information. This new surveillance/navigation capability can help provide information about what is in a wall or on the other side of a door, and can be extended to provide precise global position in areas where these services are denied such as in tunnels or caves. This paper presents our efforts along these lines including image enhancements.
{"title":"See-through-wall imaging using ultra wideband pulse systems","authors":"M. Mahfouz, A. Fathy, Yunqiang Yang, Emam ElHak Ali, A. Badawi","doi":"10.1109/AIPR.2005.40","DOIUrl":"https://doi.org/10.1109/AIPR.2005.40","url":null,"abstract":"Surveillance/navigation systems presently used make extensive use of television, infrared, and other line-of-sight-surveillance hardware. However, these systems cannot tell what is happening or locate persons/assets on the other side of a wall, behind bushes, in the dark, in a tunnel or a cave, or through a dense fog. It is our objective here to develop a new sensor, based on UWB technology. A small, lightweight, low power transceiver or multiples that are based upon the fact that microwave frequencies can be optimized to penetrate nonmetallic materials, and providing very precise ranging information. This new surveillance/navigation capability can help provide information about what is in a wall or on the other side of a door, and can be extended to provide precise global position in areas where these services are denied such as in tunnels or caves. This paper presents our efforts along these lines including image enhancements.","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133449341","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}
Girish Gopalakrishnan, S. Kumar, A. Narayanan, R. Mullick
Medical image fusion is becoming increasingly popular for enhancing diagnostic accuracy by intelligently 'fusing' information obtained from two different images. These images may be obtained from the same modality at different time instances or from multiple modalities recording complementary information. Due to the nature of the human body and also due to patient motion and breathing, there is a need for deformable registration algorithms in medical imaging. Typical nonparametric (deformable) registration algorithms such as the fluid-based, demons and curvature-based techniques are computationally intensive and have been demonstrated for mono-modality registrations only. We propose a fast and deformable algorithm using a 2-tiered strategy wherein a global MI-based affine registration is followed by a local piecewise refinement. We have tested this method on CT and PET images and validated the same using clinical experts
{"title":"A fast piecewise deformable method for multi-modality image registration","authors":"Girish Gopalakrishnan, S. Kumar, A. Narayanan, R. Mullick","doi":"10.1109/AIPR.2005.7","DOIUrl":"https://doi.org/10.1109/AIPR.2005.7","url":null,"abstract":"Medical image fusion is becoming increasingly popular for enhancing diagnostic accuracy by intelligently 'fusing' information obtained from two different images. These images may be obtained from the same modality at different time instances or from multiple modalities recording complementary information. Due to the nature of the human body and also due to patient motion and breathing, there is a need for deformable registration algorithms in medical imaging. Typical nonparametric (deformable) registration algorithms such as the fluid-based, demons and curvature-based techniques are computationally intensive and have been demonstrated for mono-modality registrations only. We propose a fast and deformable algorithm using a 2-tiered strategy wherein a global MI-based affine registration is followed by a local piecewise refinement. We have tested this method on CT and PET images and validated the same using clinical experts","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132900749","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}
The Air Force Research Laboratory Information Directorate (AFRL/IF), under sponsorship of the Department of Justice's (DOJ), National Institute of Justice (NIJ) Office of Science and Technology (OS&T), is currently developing and evaluating advanced through the wall surveillance (TWS) technologies. These technologies are partitioned into two categories: inexpensive, handheld systems for locating an individual(s) behind a wall or door; and portable, personal computer (PC) based standoff systems to enable the determination of events during critical incident situations. The technologies utilized are primarily focused on active radars operating in the UHF, L, S (ultra wideband (UWB)), X, and Ku bands. The data displayed by these systems is indicative of range (1 dimension), or range and azimuth (2 dimensions) to the moving individuals). This paper highlights the technologies employed in five (5) prototype TWS systems delivered to NIJ and AFRL/IF for test and evaluation
{"title":"An overview of through the wall surveillance for homeland security","authors":"S. Borek","doi":"10.1109/AIPR.2005.18","DOIUrl":"https://doi.org/10.1109/AIPR.2005.18","url":null,"abstract":"The Air Force Research Laboratory Information Directorate (AFRL/IF), under sponsorship of the Department of Justice's (DOJ), National Institute of Justice (NIJ) Office of Science and Technology (OS&T), is currently developing and evaluating advanced through the wall surveillance (TWS) technologies. These technologies are partitioned into two categories: inexpensive, handheld systems for locating an individual(s) behind a wall or door; and portable, personal computer (PC) based standoff systems to enable the determination of events during critical incident situations. The technologies utilized are primarily focused on active radars operating in the UHF, L, S (ultra wideband (UWB)), X, and Ku bands. The data displayed by these systems is indicative of range (1 dimension), or range and azimuth (2 dimensions) to the moving individuals). This paper highlights the technologies employed in five (5) prototype TWS systems delivered to NIJ and AFRL/IF for test and evaluation","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128663882","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}
In this paper, we analyze the perturbation in the reconstructed optical absorption images, resulting from the discretization of the forward and inverse problems. We show that the perturbation due to each problem is a function of both the forward and inverse problem solutions and can be reduced by proper refinement of the discretization mesh. Based on the perturbation analysis, we devise an adaptive discretization scheme for forward and inverse problems, which reduces the perturbation on the reconstructed image. Such a discretization scheme leads to an adaptively refined composite mesh sufficient to approximate the forward and inverse problem solutions within a desired level of accuracy while keeping the computational complexity within the computational power limits
{"title":"Discretization error based mesh generation for diffuse optical tomography","authors":"M. Guven, B. Yazıcı, Kiwoon Kwon, E. Giladi","doi":"10.1109/AIPR.2005.26","DOIUrl":"https://doi.org/10.1109/AIPR.2005.26","url":null,"abstract":"In this paper, we analyze the perturbation in the reconstructed optical absorption images, resulting from the discretization of the forward and inverse problems. We show that the perturbation due to each problem is a function of both the forward and inverse problem solutions and can be reduced by proper refinement of the discretization mesh. Based on the perturbation analysis, we devise an adaptive discretization scheme for forward and inverse problems, which reduces the perturbation on the reconstructed image. Such a discretization scheme leads to an adaptively refined composite mesh sufficient to approximate the forward and inverse problem solutions within a desired level of accuracy while keeping the computational complexity within the computational power limits","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131308441","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}
A illuminance-reflectance model based video stream enhancement algorithm is proposed for improving the visual quality of digital video streams captured by surveillance camera under insufficient and/or nonuniform lighting conditions. The paper presents computational methods for estimation of scene illuminance and reflectance, adaptive dynamic range compression of illuminance, and adaptive enhancement for mid-tone frequency components. The images are processed in a similar way as human eyes sensing a scene. The algorithm demonstrates high quality of enhanced images, robust performance and fast processing speed. Compared with Retinex and multi-scale retinex with color restoration (MSRCR), the proposed method shows a better balance between luminance enhancement and contrast enhancement as well as a more consistent and reliable color rendition without introducing incorrect colors. This is an effective technique for image enhancement with simple computational procedures, which makes real-time enhancement for homeland security application successfully realized. The application of this image enhancement technique to the FRGC images yields improved face recognition results
{"title":"An illuminance-reflectance nonlinear video enhancement model for homeland security applications","authors":"Li Tao, R. Tompkins, V. Asari","doi":"10.1109/AIPR.2005.14","DOIUrl":"https://doi.org/10.1109/AIPR.2005.14","url":null,"abstract":"A illuminance-reflectance model based video stream enhancement algorithm is proposed for improving the visual quality of digital video streams captured by surveillance camera under insufficient and/or nonuniform lighting conditions. The paper presents computational methods for estimation of scene illuminance and reflectance, adaptive dynamic range compression of illuminance, and adaptive enhancement for mid-tone frequency components. The images are processed in a similar way as human eyes sensing a scene. The algorithm demonstrates high quality of enhanced images, robust performance and fast processing speed. Compared with Retinex and multi-scale retinex with color restoration (MSRCR), the proposed method shows a better balance between luminance enhancement and contrast enhancement as well as a more consistent and reliable color rendition without introducing incorrect colors. This is an effective technique for image enhancement with simple computational procedures, which makes real-time enhancement for homeland security application successfully realized. The application of this image enhancement technique to the FRGC images yields improved face recognition results","PeriodicalId":130204,"journal":{"name":"34th Applied Imagery and Pattern Recognition Workshop (AIPR'05)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2005-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131199132","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：