Pub Date : 2009-12-04DOI: 10.1109/CCPR.2009.5343975
Cheng Cheng, Bilan Zhu, M. Nakagawa
This paper presents a revised method for keyword search from Japanese handwritten digital ink. We employ Japanese string recognition and produce a candidate lattice. We search for a given keyword into the lattice so that we can search for the keyword even if constituent characters are not in the top candidates. We present some overall performance as well as consideration on search errors.
{"title":"Ink Search Employing Japanese String Recognition","authors":"Cheng Cheng, Bilan Zhu, M. Nakagawa","doi":"10.1109/CCPR.2009.5343975","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5343975","url":null,"abstract":"This paper presents a revised method for keyword search from Japanese handwritten digital ink. We employ Japanese string recognition and produce a candidate lattice. We search for a given keyword into the lattice so that we can search for the keyword even if constituent characters are not in the top candidates. We present some overall performance as well as consideration on search errors.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"156 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":"115603329","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.2009.5344082
Tao Liu, Yue Lu
Image-based spam is becoming a new threat to the Internet and its users. In our early work, we proposed an image filtering system which detects the spam image by matching with user-specified image content using SIFT algorithm. In order to further improve efficiency, we develop a quick image matching algorithm instead of SIFT. After using difference-of-Gaussian to extract image feature points, we adopt geometry transform to judge whether two images are matched. Experimental results show that the proposed method can identify image spam without the need of OCR and it can achieve a good performance. In addition, we adopt Mean Shift algorithm to locate the highest density area of feature points, which improves the performance of the system.
{"title":"Feature Point Analysis for Image Spam E-Mail Detection","authors":"Tao Liu, Yue Lu","doi":"10.1109/CCPR.2009.5344082","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344082","url":null,"abstract":"Image-based spam is becoming a new threat to the Internet and its users. In our early work, we proposed an image filtering system which detects the spam image by matching with user-specified image content using SIFT algorithm. In order to further improve efficiency, we develop a quick image matching algorithm instead of SIFT. After using difference-of-Gaussian to extract image feature points, we adopt geometry transform to judge whether two images are matched. Experimental results show that the proposed method can identify image spam without the need of OCR and it can achieve a good performance. In addition, we adopt Mean Shift algorithm to locate the highest density area of feature points, which improves the performance of the system.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"14 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":"116989518","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}
Driving fatigue detection is a key technique in vehicle active safety. In this paper, a practical driver fatigue detection algorithm is proposed, it employs sequential detection and temporal tracking to detect human face, which combines the superiorities of both Adaboost and Mean-Shift algorithm; A morphologic filter method is given to localize the pair of eyes in the detected face area. Then multiple image features are exploited to recognize open state or close state. Various tests demonstrated that it has a performance of high detection precision and fast processing speed. To this end, it can be effectively and efficiently used in vehicle active safety systems.
{"title":"A Practical Eye State Recognition Based Driver Fatigue Detection Method","authors":"Huan Wang, Yong Cheng, Qiong Wang, Mingwu Ren, Chunxia Zhao, Jingyu Yang","doi":"10.1109/CCPR.2009.5344067","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344067","url":null,"abstract":"Driving fatigue detection is a key technique in vehicle active safety. In this paper, a practical driver fatigue detection algorithm is proposed, it employs sequential detection and temporal tracking to detect human face, which combines the superiorities of both Adaboost and Mean-Shift algorithm; A morphologic filter method is given to localize the pair of eyes in the detected face area. Then multiple image features are exploited to recognize open state or close state. Various tests demonstrated that it has a performance of high detection precision and fast processing speed. To this end, it can be effectively and efficiently used in vehicle active safety systems.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"40 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":"127205734","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.2009.5344015
Yi Zhang, Yanhua Li, Li Zeng, Q. Liu
According to the problem of the low recognition rate of speaker-independent recognition in intelligent robot, a kind of endpoint detection algorithm with double threshold is adopted and the speech endpoint can be detected accurately. The mixed parameter of Mel Frequency Cepstral Coefficients (MFCC) and fractal dimension is used as the feature parameter, and the intelligent robot command-word recognition system based on Hidden Markov Models (HMM) is realized. The recognition effect achieves above 85%. Then the performance of MFCC and the mixed parameter of MFCC and fractal dimension is contrasted and analyzed. The experiment result shows that the system recognition rate is improved by the algorithm of mixed parameter, and the system recognition performance is optimized.
{"title":"Research on Intelligent Robot Command-Word Recognition System Based on Feature Extraction","authors":"Yi Zhang, Yanhua Li, Li Zeng, Q. Liu","doi":"10.1109/CCPR.2009.5344015","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344015","url":null,"abstract":"According to the problem of the low recognition rate of speaker-independent recognition in intelligent robot, a kind of endpoint detection algorithm with double threshold is adopted and the speech endpoint can be detected accurately. The mixed parameter of Mel Frequency Cepstral Coefficients (MFCC) and fractal dimension is used as the feature parameter, and the intelligent robot command-word recognition system based on Hidden Markov Models (HMM) is realized. The recognition effect achieves above 85%. Then the performance of MFCC and the mixed parameter of MFCC and fractal dimension is contrasted and analyzed. The experiment result shows that the system recognition rate is improved by the algorithm of mixed parameter, and the system recognition performance is optimized.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"53 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":"127535347","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.2009.5344007
Xiaofeng Yang, Mingming Sun, Xuelei Hu, Jingyu Yang
HTTP-related vulnerabilities are being more commonly exploited as HTTP applications becoming the number one application across the Internet. Several HTTP specific anomaly methods have been proposed, among which grammar-based methods tend more likely to reflect the underlying structure of HTTP communications, therefore showed a promising classifying capability between benign and malicious accesses. Because of being separately proposed among other methods, grammar-based methods have not been summarized and compared directly on the same dataset. This paper presents several grammar-based anomaly methods for HTTP attacks, reveals their detecting capabilities, common features, strengths and drawbacks in comparison with each other.
{"title":"Grammar-Based Anomaly Methods for HTTP Attacks","authors":"Xiaofeng Yang, Mingming Sun, Xuelei Hu, Jingyu Yang","doi":"10.1109/CCPR.2009.5344007","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344007","url":null,"abstract":"HTTP-related vulnerabilities are being more commonly exploited as HTTP applications becoming the number one application across the Internet. Several HTTP specific anomaly methods have been proposed, among which grammar-based methods tend more likely to reflect the underlying structure of HTTP communications, therefore showed a promising classifying capability between benign and malicious accesses. Because of being separately proposed among other methods, grammar-based methods have not been summarized and compared directly on the same dataset. This paper presents several grammar-based anomaly methods for HTTP attacks, reveals their detecting capabilities, common features, strengths and drawbacks in comparison with each other.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"37 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":"125021668","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.2009.5344116
Yi Chen, Y. Bo, Ming Lv
Referring to the possible problem of gyro random drift effect on the Line of Sight(LOS) stabilization accuracy, a method for the LOS stabilization based on visual attitude estimation is proposed. A monocular vision system is established by four coplanar reference points on the ground and a double-focus CCD camera in the carrier. The quaternion is used to represent the transform relation between geodetic coordinate system and carrier coordinate system, and the square-root unscented Kalman Filter in which the current statistical model is applied to obtain the system state equations is used to estimate the attitude parameters. On this basis, the velocity interference model of the LOS is derived according to coordinate transformation and rigid body kinematics, and the method of feedforward control is adapted to isolate the influence of the LOS velocity disturbance. The simulation results indicate that the proposed method is effective in stabilizing the LOS without changing the structure of the tracking units.
{"title":"Line of Sight Stabilization Technology Based on Visual Attitude Estimation","authors":"Yi Chen, Y. Bo, Ming Lv","doi":"10.1109/CCPR.2009.5344116","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344116","url":null,"abstract":"Referring to the possible problem of gyro random drift effect on the Line of Sight(LOS) stabilization accuracy, a method for the LOS stabilization based on visual attitude estimation is proposed. A monocular vision system is established by four coplanar reference points on the ground and a double-focus CCD camera in the carrier. The quaternion is used to represent the transform relation between geodetic coordinate system and carrier coordinate system, and the square-root unscented Kalman Filter in which the current statistical model is applied to obtain the system state equations is used to estimate the attitude parameters. On this basis, the velocity interference model of the LOS is derived according to coordinate transformation and rigid body kinematics, and the method of feedforward control is adapted to isolate the influence of the LOS velocity disturbance. The simulation results indicate that the proposed method is effective in stabilizing the LOS without changing the structure of the tracking units.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"63 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":"126966276","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.2009.5343996
Wenhan Jiang, Xiaofei Zhou, Hongchuan Hou, Xinggang Lin
Support Vector Machine (SVM) needs huge computation for large scale learning tasks. Sample selection is a feasible strategy to overcome the problem. From the geometry of SVM, it is clear that a SVM problem can be converted to a problem of computing the nearest points between two convex hulls. The convex hulls virtually determine the separating plane of SVM. Since a convex hull of a set only can be constructed by boundary samples of the convex hull, using boundary samples of each class to train SVM will be equivalent to using all training samples to train the classifier. In order to select boundary samples, this paper introduces a novel sample selection strategy named Kernel Subclass Convex Hull (KSCH) sample selection strategy, which iteratively select boundary samples of each class convex hull in high dimensional space (induced by kernel trick). Experimental results on face databases show that our KSCH sample selection method can select fewer high quality samples to maintain SVM with high recognition accuracy and quickly executing speed.
{"title":"A New Sampling-Based SVM for Face Recognition","authors":"Wenhan Jiang, Xiaofei Zhou, Hongchuan Hou, Xinggang Lin","doi":"10.1109/CCPR.2009.5343996","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5343996","url":null,"abstract":"Support Vector Machine (SVM) needs huge computation for large scale learning tasks. Sample selection is a feasible strategy to overcome the problem. From the geometry of SVM, it is clear that a SVM problem can be converted to a problem of computing the nearest points between two convex hulls. The convex hulls virtually determine the separating plane of SVM. Since a convex hull of a set only can be constructed by boundary samples of the convex hull, using boundary samples of each class to train SVM will be equivalent to using all training samples to train the classifier. In order to select boundary samples, this paper introduces a novel sample selection strategy named Kernel Subclass Convex Hull (KSCH) sample selection strategy, which iteratively select boundary samples of each class convex hull in high dimensional space (induced by kernel trick). Experimental results on face databases show that our KSCH sample selection method can select fewer high quality samples to maintain SVM with high recognition accuracy and quickly executing speed.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"17 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":"114891154","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.2009.5344053
Jie Xu, Jian Yang
In this paper, an efficient feature extraction technique called Local Graph Embedding Discriminant Analysis(LGEDA) is developed for solving One Sample per Person Problem. In our algorithm, a mean filter is used to generate imitated images and a double size new training set can be obtained. Taking the local neighborhood geometry structure and class labels into account, the proposed algorithm can maximize the local interclass separability as far as possible and preserve the local neighborhood relationships of the data set. After the local scatters and interclass scatter are characterized, the proposed method seeks to find a projection maximizing the local margin between of different classes under the constraint of local neighborhood preserving. Experiments show that our proposed method
{"title":"Local Graph Embedding Discriminant Analysis for Face Recognition with Single Training Sample Per Person","authors":"Jie Xu, Jian Yang","doi":"10.1109/CCPR.2009.5344053","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344053","url":null,"abstract":"In this paper, an efficient feature extraction technique called Local Graph Embedding Discriminant Analysis(LGEDA) is developed for solving One Sample per Person Problem. In our algorithm, a mean filter is used to generate imitated images and a double size new training set can be obtained. Taking the local neighborhood geometry structure and class labels into account, the proposed algorithm can maximize the local interclass separability as far as possible and preserve the local neighborhood relationships of the data set. After the local scatters and interclass scatter are characterized, the proposed method seeks to find a projection maximizing the local margin between of different classes under the constraint of local neighborhood preserving. Experiments show that our proposed method","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"129 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":"115551463","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.2009.5344106
Lin Lin, Shijin Li, Yuelong Zhu, Lizhong Xu
To select a minimal and effective subset from a mass of bands is the key issue of the study on hyperspectral image classification. This paper put forwards a novel band selection algorithm, which combines mutual information-based grouping method and genetic algorithm. The proposed algorithm reduces the computation cost significantly, as well as keeps a better precision. In addition, resampling based on sequential clustering is employed to tackle the imbalanced data issue and improve the classification accuracy of minority classes. Experimental results on the Washington DC Mall data set validate the effectiveness and efficiency of the proposed algorithm.
{"title":"A Novel Approach to Band Selection for Hyperspectral Image Classification","authors":"Lin Lin, Shijin Li, Yuelong Zhu, Lizhong Xu","doi":"10.1109/CCPR.2009.5344106","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5344106","url":null,"abstract":"To select a minimal and effective subset from a mass of bands is the key issue of the study on hyperspectral image classification. This paper put forwards a novel band selection algorithm, which combines mutual information-based grouping method and genetic algorithm. The proposed algorithm reduces the computation cost significantly, as well as keeps a better precision. In addition, resampling based on sequential clustering is employed to tackle the imbalanced data issue and improve the classification accuracy of minority classes. Experimental results on the Washington DC Mall data set validate the effectiveness and efficiency of the proposed algorithm.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"147 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":"116086302","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.2009.5343968
Y. Hu, Benyong Liu
In this paper, a face recognition scheme is proposed, wherein face images are preprocessed based on pixel averaging and energy normalization, and features are extracted successively with fast Fourier transform and the partial least squares for dimensionality reduction, and classification results are obtained by a classifier based on hidden Markov model. Some experimental results on the Olivetti Research Laboratory face database are presented to show the feasibility of the presented scheme.
{"title":"Face Recognition Based on PLS and HMM","authors":"Y. Hu, Benyong Liu","doi":"10.1109/CCPR.2009.5343968","DOIUrl":"https://doi.org/10.1109/CCPR.2009.5343968","url":null,"abstract":"In this paper, a face recognition scheme is proposed, wherein face images are preprocessed based on pixel averaging and energy normalization, and features are extracted successively with fast Fourier transform and the partial least squares for dimensionality reduction, and classification results are obtained by a classifier based on hidden Markov model. Some experimental results on the Olivetti Research Laboratory face database are presented to show the feasibility of the presented scheme.","PeriodicalId":354468,"journal":{"name":"2009 Chinese Conference on Pattern Recognition","volume":"28 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":"114565199","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：