This paper presents a line extraction method to process images taken inside a machine tool. Instead of using real background images, our approach utilizes a virtual background image. This approach solves the problem of absence of real background images due to a dynamic background. In order to only extract lines of the object, all corners are detected from the real image first. Then, those corners generated from the background are filtered through composite background subtraction. Afterwards, a hypothesis of a line exists between any two corners is made. All the hypothetical lines are mapped back to the original real image to test for their existence. Those lines caused by noises, such as reflections or scratches, can be then eliminated. Our experimental results prove the feasibility and effectiveness of this proposed method..
{"title":"Line Extraction with Composite Background Subtract","authors":"H. Deng, X. Tian, K. Yamazaki, M. Mori","doi":"10.1109/CRV.2006.47","DOIUrl":"https://doi.org/10.1109/CRV.2006.47","url":null,"abstract":"This paper presents a line extraction method to process images taken inside a machine tool. Instead of using real background images, our approach utilizes a virtual background image. This approach solves the problem of absence of real background images due to a dynamic background. In order to only extract lines of the object, all corners are detected from the real image first. Then, those corners generated from the background are filtered through composite background subtraction. Afterwards, a hypothesis of a line exists between any two corners is made. All the hypothetical lines are mapped back to the original real image to test for their existence. Those lines caused by noises, such as reflections or scratches, can be then eliminated. Our experimental results prove the feasibility and effectiveness of this proposed method..","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114611182","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}
Image search and object recognition are two domains where it is useful to be able to describe an image in a form that is invariant to image lighting, image intensity, scaling, rotation, translation, and changes in camera position. This paper presents a method based on tracing the trajectories of interest points, specifically KLT corners, across scale-space. The KLT corner interest points are calculated with an adaptive threshold to make them invariant to image intensity. A three-dimensional point composed of two-dimensional spatial coordinates and the scale of gaussian smoothing is found for each interest point, together all the points in the image are normalized into a form that is mostly invariant to geometric changes such as scale and rotation. Each image is converted to a trajectory set which is compared between images to assess their similarity. Experiments are shown.
{"title":"Using Normalized Interest Point Trajectories Over Scale for Image Search","authors":"M. Fiala","doi":"10.1109/CRV.2006.85","DOIUrl":"https://doi.org/10.1109/CRV.2006.85","url":null,"abstract":"Image search and object recognition are two domains where it is useful to be able to describe an image in a form that is invariant to image lighting, image intensity, scaling, rotation, translation, and changes in camera position. This paper presents a method based on tracing the trajectories of interest points, specifically KLT corners, across scale-space. The KLT corner interest points are calculated with an adaptive threshold to make them invariant to image intensity. A three-dimensional point composed of two-dimensional spatial coordinates and the scale of gaussian smoothing is found for each interest point, together all the points in the image are normalized into a form that is mostly invariant to geometric changes such as scale and rotation. Each image is converted to a trajectory set which is compared between images to assess their similarity. Experiments are shown.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123956248","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}
This paper describes a new variational method for estimating disparity from stereo images. The stereo matching problem is formulated as a convex programming problem in which an objective function is minimized under various constraints modelling prior knowledge and observed information. The algorithm proposed to solve this problem has a block-iterative structure which allows a wide range of constraints to be easily incorporated, possibly taking advantage of parallel computing architectures. In this work, we use a Total Variation bound as a regularization constraint, which is shown to be well-suited to disparity maps. Experimental results for standard data sets are presented to illustrate the capabilities of the proposed disparity estimation technique.
{"title":"Disparity Map Estimation Using A Total Variation Bound","authors":"Wided Miled, J. Pesquet","doi":"10.1109/CRV.2006.28","DOIUrl":"https://doi.org/10.1109/CRV.2006.28","url":null,"abstract":"This paper describes a new variational method for estimating disparity from stereo images. The stereo matching problem is formulated as a convex programming problem in which an objective function is minimized under various constraints modelling prior knowledge and observed information. The algorithm proposed to solve this problem has a block-iterative structure which allows a wide range of constraints to be easily incorporated, possibly taking advantage of parallel computing architectures. In this work, we use a Total Variation bound as a regularization constraint, which is shown to be well-suited to disparity maps. Experimental results for standard data sets are presented to illustrate the capabilities of the proposed disparity estimation technique.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124816733","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}
While the geometric aspects of structure and motion estimation from uncalibrated images are well understood, and it has great promise in applications, it has not seen widespread use. In this paper we combine SSD tracking with incremental structure computation into a system computing both motion and structure on-line from video. We show how in combination the structure estimation and tracking benefit each other, resulting in both better structure and more robust tracking. Particularly, through the 3D structure, our method can manage visibility constraints, add new image patches to track as they come into view and remove ones that are occluded or fail. This allows tracking over larger pose variations than possible with conventional SSD tracking (e.g. going around an object or scene where new parts come into view.) Experiments demonstrate tracking and capture of a scene from a camera trajectory covering different sides without mutual visibility.
{"title":"Robust SSD tracking with incremental 3D structure estimation","authors":"Adam Rachmielowski, Dana Cobzas, Martin Jägersand","doi":"10.1109/CRV.2006.62","DOIUrl":"https://doi.org/10.1109/CRV.2006.62","url":null,"abstract":"While the geometric aspects of structure and motion estimation from uncalibrated images are well understood, and it has great promise in applications, it has not seen widespread use. In this paper we combine SSD tracking with incremental structure computation into a system computing both motion and structure on-line from video. We show how in combination the structure estimation and tracking benefit each other, resulting in both better structure and more robust tracking. Particularly, through the 3D structure, our method can manage visibility constraints, add new image patches to track as they come into view and remove ones that are occluded or fail. This allows tracking over larger pose variations than possible with conventional SSD tracking (e.g. going around an object or scene where new parts come into view.) Experiments demonstrate tracking and capture of a scene from a camera trajectory covering different sides without mutual visibility.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129067194","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}
This paper presents the status of a project targeting the development of content-based video indexing tools, to assist a human in the generation of descriptive video for the hard of seeing people. We describe three main elements: (1) the video content that is pertinent for computer-assisted descriptive video, (2) the system dataflow, based on a light plug-in architecture of an open-source video processing software and (3) the first version of the plug-ins developed to date. Plugs-ins that are under development include shot transition detection, key-frames identification, keyface detection, key-text spotting, visual motion mapping, face recognition, facial characterization, story segmentation, gait/gesture characterization, keyplace recognition, key-object spotting and image categorization. Some of these tools are adapted from our previous works on video surveillance, audiovisual speech recognition and content-based video indexing of documentary films. We do not focus on the algorithmic details in this paper neither on the global performance since the integration is done yet. We rather concentrate on discussing application issues of automatic descriptive video usability aspects.
{"title":"Toward an Application of Content-Based Video Indexing to Computer- Assisted Descriptive Video","authors":"L. Gagnon, F. Laliberté, M. Lalonde, M. Beaulieu","doi":"10.1109/CRV.2006.78","DOIUrl":"https://doi.org/10.1109/CRV.2006.78","url":null,"abstract":"This paper presents the status of a project targeting the development of content-based video indexing tools, to assist a human in the generation of descriptive video for the hard of seeing people. We describe three main elements: (1) the video content that is pertinent for computer-assisted descriptive video, (2) the system dataflow, based on a light plug-in architecture of an open-source video processing software and (3) the first version of the plug-ins developed to date. Plugs-ins that are under development include shot transition detection, key-frames identification, keyface detection, key-text spotting, visual motion mapping, face recognition, facial characterization, story segmentation, gait/gesture characterization, keyplace recognition, key-object spotting and image categorization. Some of these tools are adapted from our previous works on video surveillance, audiovisual speech recognition and content-based video indexing of documentary films. We do not focus on the algorithmic details in this paper neither on the global performance since the integration is done yet. We rather concentrate on discussing application issues of automatic descriptive video usability aspects.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126543031","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}
Conventional image-oriented cryptographic techniques lack the flexibility needed for content-specific security features such as the concealment of confidential information within a portion of a document. Content-specific security is particularly important for digital archival systems that store sensitive documents in the form of digital images. Recently, a novel image encryption scheme utilizing multiple levels of regions-of-interest (ROI) privileges for digital document encryption was developed to address the needs of modern digital document management systems. This image encryption scheme requires the selection of regions-ofinterest for encryption. The process of manually selecting regions can be time-consuming. This paper presents an automatic, regions-of-interest selection algorithm that utilizes an expert knowledge learning system to select regions of interest in a scanned document image for the purpose of minimizing human interaction time during the encryption process. Experimental results show that a high level of accuracy and significant timesaving benefits can be achieved using the proposed algorithm.
{"title":"Expert Knowledge Based Automatic Regions-of-Interest (ROI) Selection in Scanned Documents for Digital Image Encryption","authors":"A. Wong, W. Bishop","doi":"10.1109/CRV.2006.33","DOIUrl":"https://doi.org/10.1109/CRV.2006.33","url":null,"abstract":"Conventional image-oriented cryptographic techniques lack the flexibility needed for content-specific security features such as the concealment of confidential information within a portion of a document. Content-specific security is particularly important for digital archival systems that store sensitive documents in the form of digital images. Recently, a novel image encryption scheme utilizing multiple levels of regions-of-interest (ROI) privileges for digital document encryption was developed to address the needs of modern digital document management systems. This image encryption scheme requires the selection of regions-ofinterest for encryption. The process of manually selecting regions can be time-consuming. This paper presents an automatic, regions-of-interest selection algorithm that utilizes an expert knowledge learning system to select regions of interest in a scanned document image for the purpose of minimizing human interaction time during the encryption process. Experimental results show that a high level of accuracy and significant timesaving benefits can be achieved using the proposed algorithm.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"477 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129385093","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}
With the rapid growth in infrared sensor technology and its drastic cost reduction, the potential of application of these imaging technologies in computer vision systems has increased. One potential application for IR imaging is depth from stereo. It has been shown that the quality of uncooled sensors is not sufficient for generating dense depth maps. In this paper we investigate the production of sparse disparity maps for uncalibrated infrared stereo images, which necessitates a robust feature-based stereo matching technique capable of dealing with the problems of infrared images, such as low resolution and high noise. Initially, a set of stable and tractable features are extracted from stereo pairs using the phase congruency model. Then, a set of Log-Gabor wavelet coefficients in different orientations and frequencies are used to analyze and describe the extracted features for matching. Finally, epipolar geometrical constraints are employed to refine the matching results. Experiments on a set of IR stereo pairs validate the robustness of our technique.
{"title":"Sparse Disparity Map from Uncalibrated Infrared Stereo Images","authors":"K. Hajebi, J. Zelek","doi":"10.1109/CRV.2006.68","DOIUrl":"https://doi.org/10.1109/CRV.2006.68","url":null,"abstract":"With the rapid growth in infrared sensor technology and its drastic cost reduction, the potential of application of these imaging technologies in computer vision systems has increased. One potential application for IR imaging is depth from stereo. It has been shown that the quality of uncooled sensors is not sufficient for generating dense depth maps. In this paper we investigate the production of sparse disparity maps for uncalibrated infrared stereo images, which necessitates a robust feature-based stereo matching technique capable of dealing with the problems of infrared images, such as low resolution and high noise. Initially, a set of stable and tractable features are extracted from stereo pairs using the phase congruency model. Then, a set of Log-Gabor wavelet coefficients in different orientations and frequencies are used to analyze and describe the extracted features for matching. Finally, epipolar geometrical constraints are employed to refine the matching results. Experiments on a set of IR stereo pairs validate the robustness of our technique.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125266178","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 face recognition includes enhancement and segmentation of face image, detection of face boundary and facial features, matching of extracted features against the features in a database, and finally recognition of the face. The face detection algorithms are based on either gray level template matching or computation of geometric relationships among facial features. Though a number of algorithms are devised for face recognition, the technology is not matured enough to recognize the face of a person with and without beard to be the same. This research proposes a robust algorithm for preprocessing the face image with beard for face recognition. Simulation is done using eighteen JPEG images of 9 different persons having beard and without beard. This research finds application in homeland security where it can increase the robustness of the existing face recognition algorithms.
{"title":"Robust Preprocessing Algorithm for Face Recognition","authors":"G. Mittal, S. Sasi","doi":"10.1109/CRV.2006.61","DOIUrl":"https://doi.org/10.1109/CRV.2006.61","url":null,"abstract":"The face recognition includes enhancement and segmentation of face image, detection of face boundary and facial features, matching of extracted features against the features in a database, and finally recognition of the face. The face detection algorithms are based on either gray level template matching or computation of geometric relationships among facial features. Though a number of algorithms are devised for face recognition, the technology is not matured enough to recognize the face of a person with and without beard to be the same. This research proposes a robust algorithm for preprocessing the face image with beard for face recognition. Simulation is done using eighteen JPEG images of 9 different persons having beard and without beard. This research finds application in homeland security where it can increase the robustness of the existing face recognition algorithms.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"63 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125941140","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}
Previously we presented a shape reconstruction method from photometric stereo, which applies the Jacobi iterative method to reflectance map equations for M images and linearly combines the resulting iterative relations, to directly estimate the depth map of the object. For the case of two images, however; the method gives rise to noticeable distortions for certain lighting directions. In this paper, four approximations of the surface normal are introduced and the resulting 4M iterative relations are linearly combined as constraints, to effectively realize a symmetric discretization and achieve robust estimation free from such distortions. The method is investigated numerically using both synthetic and real images.
{"title":"Accurate Photometric Stereo Using Four Surface Normal Approximations","authors":"O. Ikeda","doi":"10.1109/CRV.2006.7","DOIUrl":"https://doi.org/10.1109/CRV.2006.7","url":null,"abstract":"Previously we presented a shape reconstruction method from photometric stereo, which applies the Jacobi iterative method to reflectance map equations for M images and linearly combines the resulting iterative relations, to directly estimate the depth map of the object. For the case of two images, however; the method gives rise to noticeable distortions for certain lighting directions. In this paper, four approximations of the surface normal are introduced and the resulting 4M iterative relations are linearly combined as constraints, to effectively realize a symmetric discretization and achieve robust estimation free from such distortions. The method is investigated numerically using both synthetic and real images.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122015753","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 a novel technique for face recognition is proposed. Using the statistical Local Feature Analysis (LFA) method, a set of feature points is extracted for each face image at locations with highest deviations from the expectation. Each feature point is described by a sequence of local histograms captured from the Gabor responses at different frequencies and orientations around the feature point. Histogram intersection is used to compare the Gabor histogram sequences in order to find the matched feature points between two faces. Recognition is performed based on the average similarity between the best matched points, in the probe face and each of the gallery faces. Several experiments on the FERET set of faces show the superiority of the proposed technique over all considered state-of-the-art methods (Elastic Bunch Graph Matching, LDA+PCA, Bayesian Intra/extrapersonal Classifier, Boosted Haar Classifier), and validate the robustness of our method against facial expression variation and illumination variation.
{"title":"Local Feature Matching For Face Recognition","authors":"E. F. Ersi, J. Zelek","doi":"10.1109/CRV.2006.48","DOIUrl":"https://doi.org/10.1109/CRV.2006.48","url":null,"abstract":"In this paper a novel technique for face recognition is proposed. Using the statistical Local Feature Analysis (LFA) method, a set of feature points is extracted for each face image at locations with highest deviations from the expectation. Each feature point is described by a sequence of local histograms captured from the Gabor responses at different frequencies and orientations around the feature point. Histogram intersection is used to compare the Gabor histogram sequences in order to find the matched feature points between two faces. Recognition is performed based on the average similarity between the best matched points, in the probe face and each of the gallery faces. Several experiments on the FERET set of faces show the superiority of the proposed technique over all considered state-of-the-art methods (Elastic Bunch Graph Matching, LDA+PCA, Bayesian Intra/extrapersonal Classifier, Boosted Haar Classifier), and validate the robustness of our method against facial expression variation and illumination variation.","PeriodicalId":369170,"journal":{"name":"The 3rd Canadian Conference on Computer and Robot Vision (CRV'06)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133582288","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：