Pub Date : 1992-08-30DOI: 10.1109/ICPR.1992.201780
S. Ariyoshi
Proposes a character segmentation method for Japanese printed documents. Since character segmentation is a kind of a search problem, avoiding 'combinatorial explosion' is essential in realizing practical systems. Segmentation is very complicated especially when characters touch each other. The method described gives a multi-stage algorithm, where the earlier stages treat more reliable segmentation than the later stages which utilize information obtained from the results of earlier stages. Segmentation hypotheses are generated in each stage on the basis of the results of earlier stages, and they are verified by the character recognition results. Experiments on more than one hundred documents have proven that this method is efficient and accurate for practical applications.<>
{"title":"A character segmentation method for Japanese printed documents coping with touching character problems","authors":"S. Ariyoshi","doi":"10.1109/ICPR.1992.201780","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201780","url":null,"abstract":"Proposes a character segmentation method for Japanese printed documents. Since character segmentation is a kind of a search problem, avoiding 'combinatorial explosion' is essential in realizing practical systems. Segmentation is very complicated especially when characters touch each other. The method described gives a multi-stage algorithm, where the earlier stages treat more reliable segmentation than the later stages which utilize information obtained from the results of earlier stages. Segmentation hypotheses are generated in each stage on the basis of the results of earlier stages, and they are verified by the character recognition results. Experiments on more than one hundred documents have proven that this method is efficient and accurate for practical applications.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"80 1","pages":"313-316"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74023245","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201807
Chi Hau Chen, G. You
Develops a modified probabilistic neural network (PNN) to recognize the international standard book number (ISBN). Effort is made for real time implementation of the recognition process. The authors used a simple but effective feature extraction approach, which takes the meshed averages of an isolated character as feature vector. To achieve better performance they modified the original PNN algorithm. Before normalizing the feature vectors onto a unit hyper-sphere, they are mapped into a one more dimension higher hyper-cube than the feature vector. The best result for correct characters classification is 99.62%.<>
{"title":"ISBN recognition using a modified probabilistic neural network (PNN)","authors":"Chi Hau Chen, G. You","doi":"10.1109/ICPR.1992.201807","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201807","url":null,"abstract":"Develops a modified probabilistic neural network (PNN) to recognize the international standard book number (ISBN). Effort is made for real time implementation of the recognition process. The authors used a simple but effective feature extraction approach, which takes the meshed averages of an isolated character as feature vector. To achieve better performance they modified the original PNN algorithm. Before normalizing the feature vectors onto a unit hyper-sphere, they are mapped into a one more dimension higher hyper-cube than the feature vector. The best result for correct characters classification is 99.62%.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"110 1","pages":"419-421"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79248346","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 : 1992-08-30DOI: 10.1109/ICPR.1992.202172
C. Reinhart, R. Nevatia
The authors describe the parallel implementation of a 3D object recognition algorithm. The algorithm is representative of methods utilized by various computer vision researchers and presents some interesting problems that are generally overlooked by parallel processing researchers that have studied tree search problems. They describe their objectives in developing the parallel implementation and discuss its performance. They also (briefly) discuss the affects that the parallel implementation has on the runtime characteristics of the algorithm.<>
{"title":"Issues in parallel tree search for object recognition","authors":"C. Reinhart, R. Nevatia","doi":"10.1109/ICPR.1992.202172","DOIUrl":"https://doi.org/10.1109/ICPR.1992.202172","url":null,"abstract":"The authors describe the parallel implementation of a 3D object recognition algorithm. The algorithm is representative of methods utilized by various computer vision researchers and presents some interesting problems that are generally overlooked by parallel processing researchers that have studied tree search problems. They describe their objectives in developing the parallel implementation and discuss its performance. They also (briefly) discuss the affects that the parallel implementation has on the runtime characteristics of the algorithm.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"21 1","pages":"225-228"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84974892","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201794
R. J. N. Kalberg, G. Quint, H. Scholten
Describes an interpretation system for addresses on Dutch mail. The system uses a grammar to derive the syntax of an address. After derivation of the syntax, the postal code is selected. The selection of the postal code is evaluated using a testset of 7876 images of addresses from (mainly handwritten) live mail. In 54% of the images the postal code was selected successfully. 42% of the images were rejected, in 20% rightly because the address did not contain a postal code. In 4% of the cases the algorithm misclassified a word as a postal code. The two main reasons for errors are first, the imperfect segmentation of the address into lines and words and second, the imperfect estimation of the number of characters in a word.<>
{"title":"Automatic interpretation of Dutch addresses","authors":"R. J. N. Kalberg, G. Quint, H. Scholten","doi":"10.1109/ICPR.1992.201794","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201794","url":null,"abstract":"Describes an interpretation system for addresses on Dutch mail. The system uses a grammar to derive the syntax of an address. After derivation of the syntax, the postal code is selected. The selection of the postal code is evaluated using a testset of 7876 images of addresses from (mainly handwritten) live mail. In 54% of the images the postal code was selected successfully. 42% of the images were rejected, in 20% rightly because the address did not contain a postal code. In 4% of the cases the algorithm misclassified a word as a postal code. The two main reasons for errors are first, the imperfect segmentation of the address into lines and words and second, the imperfect estimation of the number of characters in a word.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"172 1","pages":"367-370"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85018016","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201858
J. Tou
Pattern recognition studies have been traditionally concerned with concrete patterns such as two-dimensional images or graphics and three-dimensional objects. Little attention has been paid to abstract patterns such as design concepts and mathematical arguments. This paper attempts to address the abstract pattern recognition problem. The author presents an approach to automatic recognition of design concepts. Entities and relations are proposed for the description of design concepts. Computer algorithms are developed to capture the concepts in the design. The method of relaxation recognition is introduced to complete the recognition of abstract patterns.<>
{"title":"Computer recognition of design concepts","authors":"J. Tou","doi":"10.1109/ICPR.1992.201858","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201858","url":null,"abstract":"Pattern recognition studies have been traditionally concerned with concrete patterns such as two-dimensional images or graphics and three-dimensional objects. Little attention has been paid to abstract patterns such as design concepts and mathematical arguments. This paper attempts to address the abstract pattern recognition problem. The author presents an approach to automatic recognition of design concepts. Entities and relations are proposed for the description of design concepts. Computer algorithms are developed to capture the concepts in the design. The method of relaxation recognition is introduced to complete the recognition of abstract patterns.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"2 1","pages":"639-642"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82180230","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201838
H. Bunke, J. Csirik
String matching is a useful concept in pattern recognition that is constantly receiving attention from both theoretical and practical points of view. The authors propose a generalized version of the string matching algorithm by Wagner and Fischer (1974). It is based on a parametrization of the edit cost. The authors assume constant cost for any delete and insert operation, but the cost for replacing a symbol is given as a parameter r. For any two given strings A and B, the algorithm computes the edit distance of A and B in terms of the parameter r. The authors give the new algorithm and study some of its properties. Its time complexity is O(n/sup 2/.m), where n and m are the lengths of the two strings to be compared and n>
{"title":"Inference of edit costs using parametric string matching","authors":"H. Bunke, J. Csirik","doi":"10.1109/ICPR.1992.201838","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201838","url":null,"abstract":"String matching is a useful concept in pattern recognition that is constantly receiving attention from both theoretical and practical points of view. The authors propose a generalized version of the string matching algorithm by Wagner and Fischer (1974). It is based on a parametrization of the edit cost. The authors assume constant cost for any delete and insert operation, but the cost for replacing a symbol is given as a parameter r. For any two given strings A and B, the algorithm computes the edit distance of A and B in terms of the parameter r. The authors give the new algorithm and study some of its properties. Its time complexity is O(n/sup 2/.m), where n and m are the lengths of the two strings to be compared and n<or=m. The authors also discuss potential applications of the new string distance to pattern recognition. Finally, some experimental results are presented.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"21 1","pages":"549-552"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82599454","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201857
N. Bartneck, W. Ritter
An important step for image analysis is the reduction of colour levels to a small number of significant levels. This can be considered as a classification task. In this paper questions of suitable colour spaces are discussed, which have a strong correlation to the feature space used for classification. Furthermore polynomial classification as a method for colour segmentation with supervised learning is introduced. Finally results are shown coming from the application fields of traffic sign recognition and postal automation.<>
{"title":"Colour segmentation with polynomial classification","authors":"N. Bartneck, W. Ritter","doi":"10.1109/ICPR.1992.201857","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201857","url":null,"abstract":"An important step for image analysis is the reduction of colour levels to a small number of significant levels. This can be considered as a classification task. In this paper questions of suitable colour spaces are discussed, which have a strong correlation to the feature space used for classification. Furthermore polynomial classification as a method for colour segmentation with supervised learning is introduced. Finally results are shown coming from the application fields of traffic sign recognition and postal automation.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"4 1","pages":"635-638"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82663464","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201710
X. S. Cheng, E. Backer, J. J. Gerbrands
Describes a new training method, the DRBP-algorithm, for sigmoid-function based multilayer networks. The key step in DRBP is the dynamical selection and autonomous control of the learning rate. Various experiments have shown that the DRBP-algorithm has achieved its goal of fast speed, secure stability and easy parameter selection in practice.<>
{"title":"DRBP: dynamically reinforced BP-based ANN-training","authors":"X. S. Cheng, E. Backer, J. J. Gerbrands","doi":"10.1109/ICPR.1992.201710","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201710","url":null,"abstract":"Describes a new training method, the DRBP-algorithm, for sigmoid-function based multilayer networks. The key step in DRBP is the dynamical selection and autonomous control of the learning rate. Various experiments have shown that the DRBP-algorithm has achieved its goal of fast speed, secure stability and easy parameter selection in practice.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"115 1","pages":"9-12"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80363922","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201815
T. Caelli, Ashley Dreier
The authors have extended the evidence-based object recognition system of Jain and Hoffman (1988) to include some new view-independent features, a new optimized rule generation procedure based upon minimum entropy clustering and a neural network which estimates optimal evidence weights and provides an associated matching procedure. This approach provides an objective definition of the difficulty of an object recognition problem. The authors also evaluate the procedures and performance of the system with two sets of CAD (range) models.<>
{"title":"Some new techniques for evidence-based object recognition: EB-ORS1","authors":"T. Caelli, Ashley Dreier","doi":"10.1109/ICPR.1992.201815","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201815","url":null,"abstract":"The authors have extended the evidence-based object recognition system of Jain and Hoffman (1988) to include some new view-independent features, a new optimized rule generation procedure based upon minimum entropy clustering and a neural network which estimates optimal evidence weights and provides an associated matching procedure. This approach provides an objective definition of the difficulty of an object recognition problem. The authors also evaluate the procedures and performance of the system with two sets of CAD (range) models.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"31 1","pages":"450-454"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80749713","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 : 1992-08-30DOI: 10.1109/ICPR.1992.201800
K. Eom, Juha Park
A statistical contour model is developed. A digital contour is modeled by a noisy observation which is represented by polynomial functions of coordinate variables. To estimate curvature functions of digital contours, the authors develop maximum likelihood estimators by fitting the model over a small neighborhood. The neighborhood size is determined by a maximum likelihood decision rule. Statistical properties of the estimators are also investigated. The contour is decomposed at curvature extrema points by finding zero-crossings of the first derivative of estimated curvature function. Experimental results show that the model based approach performs better in estimating curvature functions and detecting extrema points than other conventional approaches based on low-pass filtered curvature functions.<>
{"title":"Contour models for curvature estimation and shape decomposition","authors":"K. Eom, Juha Park","doi":"10.1109/ICPR.1992.201800","DOIUrl":"https://doi.org/10.1109/ICPR.1992.201800","url":null,"abstract":"A statistical contour model is developed. A digital contour is modeled by a noisy observation which is represented by polynomial functions of coordinate variables. To estimate curvature functions of digital contours, the authors develop maximum likelihood estimators by fitting the model over a small neighborhood. The neighborhood size is determined by a maximum likelihood decision rule. Statistical properties of the estimators are also investigated. The contour is decomposed at curvature extrema points by finding zero-crossings of the first derivative of estimated curvature function. Experimental results show that the model based approach performs better in estimating curvature functions and detecting extrema points than other conventional approaches based on low-pass filtered curvature functions.<<ETX>>","PeriodicalId":34917,"journal":{"name":"模式识别与人工智能","volume":"19 1","pages":"393-396"},"PeriodicalIF":0.0,"publicationDate":"1992-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82331459","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：