Pub Date : 2002-12-10DOI: 10.1109/ICPR.2002.1047927
M. Silvestre, L. Ling
In this article we describe a feature extraction algorithm for pattern classification based on Bayesian decision boundaries and pruning techniques. The proposed method is capable of optimizing MLP neural classifiers by retaining those neurons in the hidden layer that really contribute to correct classification. Also, we proposed a method which defines a plausible number of neurons in the hidden layer based on the stem-and-leaf graphics of training samples. Experimental investigation reveals the efficiency of the proposed method.
{"title":"Optimization of neural classifiers based on Bayesian decision boundaries and idle neurons pruning","authors":"M. Silvestre, L. Ling","doi":"10.1109/ICPR.2002.1047927","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1047927","url":null,"abstract":"In this article we describe a feature extraction algorithm for pattern classification based on Bayesian decision boundaries and pruning techniques. The proposed method is capable of optimizing MLP neural classifiers by retaining those neurons in the hidden layer that really contribute to correct classification. Also, we proposed a method which defines a plausible number of neurons in the hidden layer based on the stem-and-leaf graphics of training samples. Experimental investigation reveals the efficiency of the proposed method.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122250292","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048227
David Tweed, A. Calway
We describe a method for tracking animals in wildlife footage. It uses a CONDENSATION particle filtering frame-work driven by learnt characteristics of specific animals. The key contribution is a periodic model of animal motion based on the relative positions over time of trackable features at significant body points. We also introduce techniques for maintaining a multimodal state density within the particle filter over time to enable consistent tracking of multiple animals. Initial experiments show that the approach has considerable potential.
{"title":"Tracking multiple animals in wildlife footage","authors":"David Tweed, A. Calway","doi":"10.1109/ICPR.2002.1048227","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048227","url":null,"abstract":"We describe a method for tracking animals in wildlife footage. It uses a CONDENSATION particle filtering frame-work driven by learnt characteristics of specific animals. The key contribution is a periodic model of animal motion based on the relative positions over time of trackable features at significant body points. We also introduce techniques for maintaining a multimodal state density within the particle filter over time to enable consistent tracking of multiple animals. Initial experiments show that the approach has considerable potential.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"160 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132444712","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048273
M. Maloof
Receiver operating characteristic (ROC) analysis is being used with greater frequency as an evaluation methodology in machine learning and pattern recognition. Researchers have used ANOVA to determine if the results from such analysis are statistically significant. Yet, in the medical decision making community, the prevailing method is LABMRMC. Although this latter method uses ANOVA, before doing so, it applies the Jackknife method to account for case-sample variance. To determine whether these two tests make the same decisions regarding statistical significance, we conducted a Monte Carlo simulation using several problems derived from Gaussian distributions, three machine-learning algorithms, ROC analysis, ANOVA, and LABMRMC. Results suggest that the decisions these tests make are not the same, even for simple problems. Furthermore, the larger issue is that since ANOVA does not account for case-sample variance, one cannot generalize experimental results to the population from which the data were drawn.
{"title":"On machine learning, ROC analysis, and statistical tests of significance","authors":"M. Maloof","doi":"10.1109/ICPR.2002.1048273","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048273","url":null,"abstract":"Receiver operating characteristic (ROC) analysis is being used with greater frequency as an evaluation methodology in machine learning and pattern recognition. Researchers have used ANOVA to determine if the results from such analysis are statistically significant. Yet, in the medical decision making community, the prevailing method is LABMRMC. Although this latter method uses ANOVA, before doing so, it applies the Jackknife method to account for case-sample variance. To determine whether these two tests make the same decisions regarding statistical significance, we conducted a Monte Carlo simulation using several problems derived from Gaussian distributions, three machine-learning algorithms, ROC analysis, ANOVA, and LABMRMC. Results suggest that the decisions these tests make are not the same, even for simple problems. Furthermore, the larger issue is that since ANOVA does not account for case-sample variance, one cannot generalize experimental results to the population from which the data were drawn.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131178041","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048280
Xiang-Yan Zeng, Yenwei Chen, Z. Nakao
In subspace pattern recognition, the basis vectors represent the features of the data and define the class. In the previous works, the standard principal component analysis is used to derive the basis vectors. Compared with the standard PCA, a nonlinear PCA can provide the high-order statistics and result in non-orthogonal basis vectors. We combine a nonlinear PCA and a subspace classifier to extract the edge and line features in an image. The simulation results indicate that the basis vectors from the nonlinear PCA can classify the edge patterns better than those from a linear PCA.
{"title":"Image feature representation by the subspace of nonlinear PCA","authors":"Xiang-Yan Zeng, Yenwei Chen, Z. Nakao","doi":"10.1109/ICPR.2002.1048280","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048280","url":null,"abstract":"In subspace pattern recognition, the basis vectors represent the features of the data and define the class. In the previous works, the standard principal component analysis is used to derive the basis vectors. Compared with the standard PCA, a nonlinear PCA can provide the high-order statistics and result in non-orthogonal basis vectors. We combine a nonlinear PCA and a subspace classifier to extract the edge and line features in an image. The simulation results indicate that the basis vectors from the nonlinear PCA can classify the edge patterns better than those from a linear PCA.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133427726","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048440
Jiun-Hung Chen, Chu-Song Chen
In this paper, we propose a new method to speed up SVM decision based on the idea of mirror points. Decisions based on multiple simple classifiers, which are formed as a result of mirror pairs, are combined to approximate a single SVM. A dynamic programming-based method is used to find a suitable combination. Experimental results show that this method can increase classification efficiencies of SVM with comparable classification performances.
{"title":"Speeding up SVM decision based on mirror points","authors":"Jiun-Hung Chen, Chu-Song Chen","doi":"10.1109/ICPR.2002.1048440","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048440","url":null,"abstract":"In this paper, we propose a new method to speed up SVM decision based on the idea of mirror points. Decisions based on multiple simple classifiers, which are formed as a result of mirror pairs, are combined to approximate a single SVM. A dynamic programming-based method is used to find a suitable combination. Experimental results show that this method can increase classification efficiencies of SVM with comparable classification performances.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114129877","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048475
K. Messer, W. Christmas, J. Kittler
ASSAVID is an EU sponsored project which is concerned with the development of a system for the automatic segmentation and semantic annotation of sports video. We describe the method which automatically classifies unknown sports video into the type of sport being played. This is an important task if a fully automatic sports video logging process is to be realised. The proposed technique relies upon the concept of "cues" which attach semantic meaning to low-level features computed on the video. Experimental results on sports video provided by the BBC demonstrate that this method is working well.
{"title":"Automatic sports classification","authors":"K. Messer, W. Christmas, J. Kittler","doi":"10.1109/ICPR.2002.1048475","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048475","url":null,"abstract":"ASSAVID is an EU sponsored project which is concerned with the development of a system for the automatic segmentation and semantic annotation of sports video. We describe the method which automatically classifies unknown sports video into the type of sport being played. This is an important task if a fully automatic sports video logging process is to be realised. The proposed technique relies upon the concept of \"cues\" which attach semantic meaning to low-level features computed on the video. Experimental results on sports video provided by the BBC demonstrate that this method is working well.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114589325","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048457
A. Sappa
This paper presents a process to improve the quality of range image segmentation by using geometrical relationships. The proposed technique consists of studying the surface continuity of an automatically generated surface model. Generally, surfaces are extracted independently (e.g., by means of a region growing algorithm) thus information about their connectivity is lost. Assuming that in most of the cases a surface cannot be disconnected with the others present in the given scene, occluded areas and crease edges can be recovered. Occluded regions are recovered by connecting surfaces that are represented by the same parameters. In addition, enforcing geometrical constraints, such as surface intersections, crease edges are recovered improving significantly the final model. Experimental results with automatically segmented real range images are presented.
{"title":"Improving segmentation results by studying surface continuity","authors":"A. Sappa","doi":"10.1109/ICPR.2002.1048457","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048457","url":null,"abstract":"This paper presents a process to improve the quality of range image segmentation by using geometrical relationships. The proposed technique consists of studying the surface continuity of an automatically generated surface model. Generally, surfaces are extracted independently (e.g., by means of a region growing algorithm) thus information about their connectivity is lost. Assuming that in most of the cases a surface cannot be disconnected with the others present in the given scene, occluded areas and crease edges can be recovered. Occluded regions are recovered by connecting surfaces that are represented by the same parameters. In addition, enforcing geometrical constraints, such as surface intersections, crease edges are recovered improving significantly the final model. Experimental results with automatically segmented real range images are presented.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128281369","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048224
Yong Zhang, Dmitry Goldgof, Sudeep Sarkar, L. Tsap
This paper presents a physical model-based method for recovering and tracking nonrigid motion of elastic objects. The proposed method recovers the motion in terms of actual physical parameters (Young's modulus) that characterize the dynamics of the objects. The tracking scheme synthesizes the motion of the points inside the object from the boundary observations, constrained by the physical parameters. Experiments on three image sequences show that using the recovered physical parameters as constraints can greatly improve the tracking quality.
{"title":"Tracking objects using recovered physical motion parameters","authors":"Yong Zhang, Dmitry Goldgof, Sudeep Sarkar, L. Tsap","doi":"10.1109/ICPR.2002.1048224","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048224","url":null,"abstract":"This paper presents a physical model-based method for recovering and tracking nonrigid motion of elastic objects. The proposed method recovers the motion in terms of actual physical parameters (Young's modulus) that characterize the dynamics of the objects. The tracking scheme synthesizes the motion of the points inside the object from the boundary observations, constrained by the physical parameters. Experiments on three image sequences show that using the recovered physical parameters as constraints can greatly improve the tracking quality.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128672733","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048349
A. Minagawa, K. Uda, N. Tagawa
We show a fast algorithm for region extraction based on belief propagation with loopy networks. The solution to this region segmentation problem, which includes the region extraction problem, is of significant computational cost if a conventional iterative approach or statistical sampling methods are applied. In the proposed approach, Gaussian loopy belief propagation is applied to a continuous-valued problem that replaces the discrete labeling problem. We show that the computational cost for region extraction can be reduced by using this algorithm, and apply the method to the extraction of a discontinuous area in Moire topography.
{"title":"Region extraction based on belief propagation for gaussian model","authors":"A. Minagawa, K. Uda, N. Tagawa","doi":"10.1109/ICPR.2002.1048349","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048349","url":null,"abstract":"We show a fast algorithm for region extraction based on belief propagation with loopy networks. The solution to this region segmentation problem, which includes the region extraction problem, is of significant computational cost if a conventional iterative approach or statistical sampling methods are applied. In the proposed approach, Gaussian loopy belief propagation is applied to a continuous-valued problem that replaces the discrete labeling problem. We show that the computational cost for region extraction can be reduced by using this algorithm, and apply the method to the extraction of a discontinuous area in Moire topography.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134228343","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 : 2002-12-10DOI: 10.1109/ICPR.2002.1048472
Chi Hau Chen, U. Qidwai
Ultrasonic imaging for nondestructive evaluation (NDE) applications is an important process for industrial applications. Such images are constrained by the sensor positions and indirect image formation. In this paper, some recent techniques in the areas of ultrasonic image enhancement and restoration, developed by the authors, are presented. Three new approaches have been presented to enhance the ultrasonic images with minimum or no information of the distortion function or the imaging system characteristics.
{"title":"Recent trends in 2D blind deconvolution for nondestructive evaluation","authors":"Chi Hau Chen, U. Qidwai","doi":"10.1109/ICPR.2002.1048472","DOIUrl":"https://doi.org/10.1109/ICPR.2002.1048472","url":null,"abstract":"Ultrasonic imaging for nondestructive evaluation (NDE) applications is an important process for industrial applications. Such images are constrained by the sensor positions and indirect image formation. In this paper, some recent techniques in the areas of ultrasonic image enhancement and restoration, developed by the authors, are presented. Three new approaches have been presented to enhance the ultrasonic images with minimum or no information of the distortion function or the imaging system characteristics.","PeriodicalId":159502,"journal":{"name":"Object recognition supported by user interaction for service robots","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2002-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134416765","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：