Pub Date : 2009-12-04DOI: 10.1109/CCPR.2009.5343964
Hui Xu, Y. Qu, Yan Zhang, Feng Zhao
A critical step in fingerprint recognition is to skeletonize the fingerprint image for minutiae extraction, which is recognized as "thinning" in image processing. The speed and reliability of the thinning process are important for the whole fingerprint identification system. In this paper, to accelerate the thinning process, a fast hardware thinning algorithm is implemented on the Xilinx Virtex II Pro developing system with a highly- paralleled architecture. Appealing experimental result is presented and the advantage of hardware thinning is also explored.
指纹识别的一个关键步骤是对指纹图像进行骨架化以提取细节，这在图像处理中被称为“细化”。细化过程的速度和可靠性对整个指纹识别系统至关重要。为了加速细化过程，本文在高并行架构的Xilinx Virtex II Pro开发系统上实现了一种快速硬件细化算法。给出了令人满意的实验结果，并探讨了硬件细化的优点。
{"title":"FPGA Based Parallel Thinning for Binary Fingerprint Image","authors":"Hui Xu, Y. Qu, Yan Zhang, Feng Zhao","doi":"10.1109/CCPR.2009.5343964","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5343964","url":null,"abstract":"A critical step in fingerprint recognition is to skeletonize the fingerprint image for minutiae extraction, which is recognized as \"thinning\" in image processing. The speed and reliability of the thinning process are important for the whole fingerprint identification system. In this paper, to accelerate the thinning process, a fast hardware thinning algorithm is implemented on the Xilinx Virtex II Pro developing system with a highly- paralleled architecture. Appealing experimental result is presented and the advantage of hardware thinning is also explored.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121247176","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344118
Yunyun Cao, F. Da, Y. Sui
In order to solve the well-known streaking effects of dynamic programming, an improved algorithm based on stereo matching technology is proposed to generate 3D building model. This algorithm obtains the feature points of the building by Harris corner detector and Canny edge detector to segment the scan lines of dynamic programming. Moreover, a linearly interpolated dissimilarity measure is introduced into the cost computation which further improves the matching speed. The experimental results show that the proposed algorithm can produce smooth and dense 3D points cloud model of building.
{"title":"A Stereo Matching Based 3D Building Reconstruction Algorithm","authors":"Yunyun Cao, F. Da, Y. Sui","doi":"10.1109/CCPR.2009.5344118","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344118","url":null,"abstract":"In order to solve the well-known streaking effects of dynamic programming, an improved algorithm based on stereo matching technology is proposed to generate 3D building model. This algorithm obtains the feature points of the building by Harris corner detector and Canny edge detector to segment the scan lines of dynamic programming. Moreover, a linearly interpolated dissimilarity measure is introduced into the cost computation which further improves the matching speed. The experimental results show that the proposed algorithm can produce smooth and dense 3D points cloud model of building.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127027096","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5343952
Tomo Miyazaki, S. Omachi, H. Aso
Extracting structure information from decorative character images is a challenging problem in the field of character recognition. The structure information of a decorative character image can be represented by a graph. However, the topologies of graphs are different even if they are the ones of the same character, because of various decorations. In this paper, we propose a method to extract a representative graph of decorative character images. The proposed method extracts graphs from decorative character images, obtains common nodes in a character and iteratively integrates graphs into one common supergraph using common nodes. To show the validly of the proposed method, experiments are carried out using decorative character images.
{"title":"Extraction of Representative Structure of Decorative Character Images","authors":"Tomo Miyazaki, S. Omachi, H. Aso","doi":"10.1109/CCPR.2009.5343952","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5343952","url":null,"abstract":"Extracting structure information from decorative character images is a challenging problem in the field of character recognition. The structure information of a decorative character image can be represented by a graph. However, the topologies of graphs are different even if they are the ones of the same character, because of various decorations. In this paper, we propose a method to extract a representative graph of decorative character images. The proposed method extracts graphs from decorative character images, obtains common nodes in a character and iteratively integrates graphs into one common supergraph using common nodes. To show the validly of the proposed method, experiments are carried out using decorative character images.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126531522","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344155
Jie-sheng Wang, Xian-wen Gao
Aiming at the predifined clustering number, strong randomness and easiness to fall into local optimum , a new self-adaptive FCM algorithm based on genetic algorithm is proposed. The number of fuzzy clustering and cluster centers are optimized by sizable-chromosome genetic algorithms (SC-GAs). Cut operator and splice operator are adopted to combination the chromosome to form new individuals. Non-uniform mutation operator is used to enhance the population diversity. The new proposed method can obtain the global optimam compared to standard FCM algorithm. The simulation experimental result s with IRIS demonstrate the feasibility and effectiveness of the new algorithm.
{"title":"Optimization of Fuzzy C-Means Clustering by Genetic Algorithms Based on Sizable Chromosome","authors":"Jie-sheng Wang, Xian-wen Gao","doi":"10.1109/CCPR.2009.5344155","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344155","url":null,"abstract":"Aiming at the predifined clustering number, strong randomness and easiness to fall into local optimum , a new self-adaptive FCM algorithm based on genetic algorithm is proposed. The number of fuzzy clustering and cluster centers are optimized by sizable-chromosome genetic algorithms (SC-GAs). Cut operator and splice operator are adopted to combination the chromosome to form new individuals. Non-uniform mutation operator is used to enhance the population diversity. The new proposed method can obtain the global optimam compared to standard FCM algorithm. The simulation experimental result s with IRIS demonstrate the feasibility and effectiveness of the new algorithm.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132397530","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344021
Xu Pan, Hong-Bin Gu, Chanjuan Sun
In this paper, we introduce a classification approach to identify definitions of all terms from a aviation professional corpus. The corpora of aviation domain are firstly segmented by LTP platform from HIT. Then four feature selection methods and two classifiers are applied to extract definitions. First of all, we summarize the correct proportion of feature subset used in classification of term definitions, and secondly argue that the naive bayes classifier combined with CHI or ODDS for feature selection achieve the best score in the F1-measure and F2-measure. In the end, we recognize that the use of SVM classifier with linear kernel could achieve very high precision, but the worst recall.
{"title":"A Classification Approach to Identify Definitions in Aviation Domain","authors":"Xu Pan, Hong-Bin Gu, Chanjuan Sun","doi":"10.1109/CCPR.2009.5344021","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344021","url":null,"abstract":"In this paper, we introduce a classification approach to identify definitions of all terms from a aviation professional corpus. The corpora of aviation domain are firstly segmented by LTP platform from HIT. Then four feature selection methods and two classifiers are applied to extract definitions. First of all, we summarize the correct proportion of feature subset used in classification of term definitions, and secondly argue that the naive bayes classifier combined with CHI or ODDS for feature selection achieve the best score in the F1-measure and F2-measure. In the end, we recognize that the use of SVM classifier with linear kernel could achieve very high precision, but the worst recall.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"73 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130736971","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344137
Shuli Han, Bo Yuan, Wenhuang Liu
Rare class problems exist extensively in real-world applications across a wide range of domains. The extreme scarcity of the target class challenges traditional machine learning algorithms focusing on the overall classification accuracy. As a result, purposefully designed techniques are required for effectively solving the rare class mining problem. This paper presents a systematic review of the major representative approaches to rare class mining and related topics and gives a summary of the important research directions.
{"title":"Rare Class Mining: Progress and Prospect","authors":"Shuli Han, Bo Yuan, Wenhuang Liu","doi":"10.1109/CCPR.2009.5344137","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344137","url":null,"abstract":"Rare class problems exist extensively in real-world applications across a wide range of domains. The extreme scarcity of the target class challenges traditional machine learning algorithms focusing on the overall classification accuracy. As a result, purposefully designed techniques are required for effectively solving the rare class mining problem. This paper presents a systematic review of the major representative approaches to rare class mining and related topics and gives a summary of the important research directions.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"90 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132511212","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344101
Jian-hua Yuan
The super-resolution image reconstruction is an ill-posed problem, which need regularizing during the reconstruction. The super-resolution image was modeled a two-dimensional manifold embedded in a three-dimensional space. The regularization constraint in the reconstruction was that the image was the minimal surface on the two-dimensional manifold. The algorithm broadened the image restoration algorithms based on the partial differential equation, and the TV restoration algorithm was a particular case of the minimal surface constraint reconstruction algorithm. The experiments show the algorithm could reconstruct the super-resolution image efficiently.
{"title":"Super-Resolution Image Reconstruction Based on the Minimal Surface Constraint on the Manifold","authors":"Jian-hua Yuan","doi":"10.1109/CCPR.2009.5344101","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344101","url":null,"abstract":"The super-resolution image reconstruction is an ill-posed problem, which need regularizing during the reconstruction. The super-resolution image was modeled a two-dimensional manifold embedded in a three-dimensional space. The regularization constraint in the reconstruction was that the image was the minimal surface on the two-dimensional manifold. The algorithm broadened the image restoration algorithms based on the partial differential equation, and the TV restoration algorithm was a particular case of the minimal surface constraint reconstruction algorithm. The experiments show the algorithm could reconstruct the super-resolution image efficiently.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"162 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134278236","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344128
Zhenghong Gu, Jian Yang
Marginal information is of great importance for classification. This paper presents a new nonparametric linear discriminant analysis method named Push-Pull marginal discriminant analysis (PPMDA) which takes full advantage of marginal information. For two-class cases, the idea of this method is to determine projection directions such that the marginal samples of one class are pushed away from the between-class marginal samples as far as possible and simultaneously pulled to the within-class samples as close as possible. This idea can be extended for multi-class cases and gives rise to the PPMDA algorithm for feature extraction of multi-class problems. The proposed method is evaluated using the Extended Yale face database B and the ORL database. Experimental results show the effectiveness of the proposed method and its performance advantage over the state-of-art feature extraction methods
{"title":"A New Nonparametric Linear Discriminant Analysis Method Based on Marginal Information","authors":"Zhenghong Gu, Jian Yang","doi":"10.1109/CCPR.2009.5344128","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344128","url":null,"abstract":"Marginal information is of great importance for classification. This paper presents a new nonparametric linear discriminant analysis method named Push-Pull marginal discriminant analysis (PPMDA) which takes full advantage of marginal information. For two-class cases, the idea of this method is to determine projection directions such that the marginal samples of one class are pushed away from the between-class marginal samples as far as possible and simultaneously pulled to the within-class samples as close as possible. This idea can be extended for multi-class cases and gives rise to the PPMDA algorithm for feature extraction of multi-class problems. The proposed method is evaluated using the Extended Yale face database B and the ORL database. Experimental results show the effectiveness of the proposed method and its performance advantage over the state-of-art feature extraction methods","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"112 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133962556","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344023
Lili Ma, Dongfeng Cai, Lanhai Zhou, Na Ye
This paper proposes a method which is aimed to translate English patent terms into Chinese based on head-driven method. Firstly, word alignment information and English NP parse tree are formed. The corresponding relation between word alignment information and syntactic structure which is built using restrict of head. The NP translation pattern database is formed as the gist of term reordering. Then the intermediate result is translated using statistical method. The best result is chose according to mutual information between each modifier and head. Experimental results show the significant improvements over the current phrase-base SMT system.
{"title":"Term Translation Based on Head-Driven Method","authors":"Lili Ma, Dongfeng Cai, Lanhai Zhou, Na Ye","doi":"10.1109/CCPR.2009.5344023","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344023","url":null,"abstract":"This paper proposes a method which is aimed to translate English patent terms into Chinese based on head-driven method. Firstly, word alignment information and English NP parse tree are formed. The corresponding relation between word alignment information and syntactic structure which is built using restrict of head. The NP translation pattern database is formed as the gist of term reordering. Then the intermediate result is translated using statistical method. The best result is chose according to mutual information between each modifier and head. Experimental results show the significant improvements over the current phrase-base SMT system.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127749077","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 : 2009-12-04DOI: 10.1109/CCPR.2009.5344062
Tong Luo, Yuquan Chen, Jianfeng Li, Jianhua Li
Typical low level segmentation method like level set method can be explained in maximum a posteriori estimation (MAP) for pixel label. In this paper, CRFs model is introduced in label estimation combined with level set to produce fast low level process and accurate high level inference. The energy term in level set evolution is also extended to contain object spatial factors, gradient is provided as the spatial updating basis, besides the temporal characteristic in curve evolution. Unlike simple CRFs model, a feedback machinery is imported in parameters learning, the reasons lie in the fact that CRFs could has small sample size and its modeling approach is mainly rely on model structure, but image patch is a typical local feature which is not directly applied into. With image patch used in the feedback, the accuracy of learning can be improved. At last, energy function is extended to allow complicated multiple regions competition, the local features is merged in the process.
{"title":"Dynamic Estimation of Curve Evolution in Image Segmentation with CRFs Label Inferring","authors":"Tong Luo, Yuquan Chen, Jianfeng Li, Jianhua Li","doi":"10.1109/CCPR.2009.5344062","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344062","url":null,"abstract":"Typical low level segmentation method like level set method can be explained in maximum a posteriori estimation (MAP) for pixel label. In this paper, CRFs model is introduced in label estimation combined with level set to produce fast low level process and accurate high level inference. The energy term in level set evolution is also extended to contain object spatial factors, gradient is provided as the spatial updating basis, besides the temporal characteristic in curve evolution. Unlike simple CRFs model, a feedback machinery is imported in parameters learning, the reasons lie in the fact that CRFs could has small sample size and its modeling approach is mainly rely on model structure, but image patch is a typical local feature which is not directly applied into. With image patch used in the feedback, the accuracy of learning can be improved. At last, energy function is extended to allow complicated multiple regions competition, the local features is merged in the process.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2009-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116957659","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：