Audio classification is based on audio features. The choice of audio features can reflect important audio classification features in time and frequency time. The extraction and analysis of audio features are the base and important of audio classification. The most important problem is to extract audio features effectively and make them mutual independence to reduce information redundancy. In this paper, combined with independent component analysis and rough set, a method for audio feature extraction is presented and it's proved better performance by experiments.
{"title":"Hybrid Independent Component Analysis and Rough Set Approach for Audio Feature Extraction","authors":"Xin He, Ling Guo, Jianyu Wang, Xianzhong Zhou","doi":"10.1109/CCPR.2008.86","DOIUrl":"https://doi.org/10.1109/CCPR.2008.86","url":null,"abstract":"Audio classification is based on audio features. The choice of audio features can reflect important audio classification features in time and frequency time. The extraction and analysis of audio features are the base and important of audio classification. The most important problem is to extract audio features effectively and make them mutual independence to reduce information redundancy. In this paper, combined with independent component analysis and rough set, a method for audio feature extraction is presented and it's proved better performance by experiments.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133797333","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 rapidly development of the biometrics market, the importance of quality control of the products should be more and more highlighted. And the evaluation of the biometrics algorithm which is the essential part of the products turned to be particularly important. On this background and according to the requirement of national research item, an evaluation scheme and system which is characterized as multi-mode Windows/Linux OS-based and automatic has been designed and developed. Here will be a introduction of the system on this paper.
{"title":"A Multi-Mode Windows/Linux OS-Based Evaluation Scheme and System of Biometrics Algorithm","authors":"Zhen Li, Zhenan Sun, Yong Zou, T. Tan","doi":"10.1109/CCPR.2008.63","DOIUrl":"https://doi.org/10.1109/CCPR.2008.63","url":null,"abstract":"With the rapidly development of the biometrics market, the importance of quality control of the products should be more and more highlighted. And the evaluation of the biometrics algorithm which is the essential part of the products turned to be particularly important. On this background and according to the requirement of national research item, an evaluation scheme and system which is characterized as multi-mode Windows/Linux OS-based and automatic has been designed and developed. Here will be a introduction of the system on this paper.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129370755","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}
To make the moving object detection faster and more reliable, in this paper we present a novel method based on fast approximated SIFT descriptor. The main idea is to compute the feature descriptor of a key-point using the integral histogram of the surrounding squared region. The feature descriptor could be further used in the feature matching between two sequential frames in the image sequence. When involved in calculating hundreds of feature descriptors, this method is profitable as it reduced computational cost, accelerated the computational speed while still maintained a fairly stable matching performance compared with the traditional SIFT descriptor. The experimental results showed that it was nearly three times faster than before and was able to meet more restrict real-time requirements.
{"title":"Fast Approximated SIFT Applied in Moving Objects Detection","authors":"Wei Tang, Zhaoshun Wang","doi":"10.1109/CCPR.2008.47","DOIUrl":"https://doi.org/10.1109/CCPR.2008.47","url":null,"abstract":"To make the moving object detection faster and more reliable, in this paper we present a novel method based on fast approximated SIFT descriptor. The main idea is to compute the feature descriptor of a key-point using the integral histogram of the surrounding squared region. The feature descriptor could be further used in the feature matching between two sequential frames in the image sequence. When involved in calculating hundreds of feature descriptors, this method is profitable as it reduced computational cost, accelerated the computational speed while still maintained a fairly stable matching performance compared with the traditional SIFT descriptor. The experimental results showed that it was nearly three times faster than before and was able to meet more restrict real-time requirements.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"155 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123609632","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}
An improved algorithm based on stereo matching technology is proposed to generate 3D face model. By improving and combining the region growing and dynamic programming methods together, a new algorithm of stereo matching is proposed. By processing region growing and dynamic programming methods on two nearly perpendicular directions separately, this new algorithm implements smooth restriction on multi-directions along with effective control of processing direction, which has improved the accuracy of matching notably. Besides, by choosing the outlines of face features as the baselines of region growing, the information of face characters also assists matching. This further improves the accuracy. Experimental results show that the reconstruction results are smooth and vivid.
{"title":"A Stereo Matching based 3D Face Reconstruction Algorithm","authors":"Youcheng Fu, F. Da","doi":"10.1109/CCPR.2008.57","DOIUrl":"https://doi.org/10.1109/CCPR.2008.57","url":null,"abstract":"An improved algorithm based on stereo matching technology is proposed to generate 3D face model. By improving and combining the region growing and dynamic programming methods together, a new algorithm of stereo matching is proposed. By processing region growing and dynamic programming methods on two nearly perpendicular directions separately, this new algorithm implements smooth restriction on multi-directions along with effective control of processing direction, which has improved the accuracy of matching notably. Besides, by choosing the outlines of face features as the baselines of region growing, the information of face characters also assists matching. This further improves the accuracy. Experimental results show that the reconstruction results are smooth and vivid.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128803749","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}
KPCA based pattern denoising has been addressed. This method, based on machine learning, maps nonlinearly patterns in input space into a higher-dimensional feature space by kernel functions, then performs PCA in feature space to realize pattern denoising. The key difficulty for this method is to seek the pre-image or an approximate pre-image in input space corresponding to the pattern after denoising in feature space. This paper proposes to utilize particle swarm optimization (PSO) algorithms to find pre-images in input space. Some nearest training patterns from the pre-image are selected as the initial group of PSO, then PSO algorithm performs an iterative process to find the pre-image or a best approximate pre-image. Experimental results based on the USPS dataset show that our proposed method outperforms some traditional techniques. Additionally, the PSO-based method is straightforward to understand, and is also easy to realize.
{"title":"Combining KPCA and PSO for Pattern Denoising","authors":"Jianwu Li, Lu Su","doi":"10.1109/CCPR.2008.10","DOIUrl":"https://doi.org/10.1109/CCPR.2008.10","url":null,"abstract":"KPCA based pattern denoising has been addressed. This method, based on machine learning, maps nonlinearly patterns in input space into a higher-dimensional feature space by kernel functions, then performs PCA in feature space to realize pattern denoising. The key difficulty for this method is to seek the pre-image or an approximate pre-image in input space corresponding to the pattern after denoising in feature space. This paper proposes to utilize particle swarm optimization (PSO) algorithms to find pre-images in input space. Some nearest training patterns from the pre-image are selected as the initial group of PSO, then PSO algorithm performs an iterative process to find the pre-image or a best approximate pre-image. Experimental results based on the USPS dataset show that our proposed method outperforms some traditional techniques. Additionally, the PSO-based method is straightforward to understand, and is also easy to realize.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126401297","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}
Determining, or selecting a distance measure over the input feature space is a fundamental problem in pattern recognition. A notable metric, called the image euclidean distance (IMED) was proposed by Wang et al. [5], which is demonstrated consistent performance improvements in many real-world problems. In this paper, we present a fast implementation of IMED, which is referred as the convolution standardizing transform (CST). It can reduce the space complexity from O(n12n22 ) to O(1) , and the time complexity from O(n12n22 ) to O(n1n2), for n1 X n2 images. Both theoretical analysis and experimental results show the efficiency of our algorithm.
{"title":"A Fast Algorithm for Image Euclidean Distance","authors":"Bing Sun, Jufu Feng","doi":"10.1109/CCPR.2008.32","DOIUrl":"https://doi.org/10.1109/CCPR.2008.32","url":null,"abstract":"Determining, or selecting a distance measure over the input feature space is a fundamental problem in pattern recognition. A notable metric, called the image euclidean distance (IMED) was proposed by Wang et al. [5], which is demonstrated consistent performance improvements in many real-world problems. In this paper, we present a fast implementation of IMED, which is referred as the convolution standardizing transform (CST). It can reduce the space complexity from O(n<sub>1</sub> <sup>2</sup>n<sub>2</sub> <sup>2</sup> ) to O(1) , and the time complexity from O(n<sub>1</sub> <sup>2</sup>n<sub>2</sub> <sup>2</sup> ) to O(n<sub>1</sub>n<sub>2</sub>), for n<sub>1</sub> X n<sub>2</sub> images. Both theoretical analysis and experimental results show the efficiency of our algorithm.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131130849","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 proposed a blind digital image watermark analysis algorithm, which is to analyze watermarks in host images without any prior-watermarking embedding and detection information. In the proposed algorithm, DWT is firstly used to decompose images into detail subbands, and noise visibility function is used to enhance the detail subbands, then non-negative matrix factorization (NMF) is used to reveal the intrinsic features in host images. Support vector machines (SVMs) are finally used to classify these characteristics. Numerical experimental results show that the proposed scheme describes the intrinsic statistical characteristics and the proposed watermark analysis is effective.
{"title":"Blind Image Watermark Analysis using DWT and Non-Negative Matrix Factorization","authors":"Wei Sun, Wei Lu","doi":"10.1109/CCPR.2008.71","DOIUrl":"https://doi.org/10.1109/CCPR.2008.71","url":null,"abstract":"This paper proposed a blind digital image watermark analysis algorithm, which is to analyze watermarks in host images without any prior-watermarking embedding and detection information. In the proposed algorithm, DWT is firstly used to decompose images into detail subbands, and noise visibility function is used to enhance the detail subbands, then non-negative matrix factorization (NMF) is used to reveal the intrinsic features in host images. Support vector machines (SVMs) are finally used to classify these characteristics. Numerical experimental results show that the proposed scheme describes the intrinsic statistical characteristics and the proposed watermark analysis is effective.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129356290","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 we propose a novel data driven strategy for designing Gabor wavelets for face recognition. Each face image is represented through a multi-sensor scheme, which splits the 2D frequency plane into a number of channels and identifies the most significant units for extracting information. The representative units for a set of face images are then derived based on statistical analysis of these units. The locations of these units in the 2D frequency plane are then used to design the frequency and orientation of Gabor wavelets for face recognition. Once frequency and orientation are determined, the scale of a Gabor wavelet is determined by the sharpness of the filtered images. Two Gabor wavelet based face recognition algorithms are applied to demonstrate the advantages of the proposed strategy against conventional parameter settings. Experimental results show that the face recognition algorithms using the designed Gabor wavelets achieve better performance in terms of accuracy and efficiency. Since the strategy is based on the training data, it can be easily applied to designing Gabor wavelets for general pattern recognition task.
{"title":"Data Driven Gabor Wavelet Design for Face Recognition","authors":"L. Shen, L. Bai, Z. Ji","doi":"10.1109/CCPR.2008.55","DOIUrl":"https://doi.org/10.1109/CCPR.2008.55","url":null,"abstract":"In this paper we propose a novel data driven strategy for designing Gabor wavelets for face recognition. Each face image is represented through a multi-sensor scheme, which splits the 2D frequency plane into a number of channels and identifies the most significant units for extracting information. The representative units for a set of face images are then derived based on statistical analysis of these units. The locations of these units in the 2D frequency plane are then used to design the frequency and orientation of Gabor wavelets for face recognition. Once frequency and orientation are determined, the scale of a Gabor wavelet is determined by the sharpness of the filtered images. Two Gabor wavelet based face recognition algorithms are applied to demonstrate the advantages of the proposed strategy against conventional parameter settings. Experimental results show that the face recognition algorithms using the designed Gabor wavelets achieve better performance in terms of accuracy and efficiency. Since the strategy is based on the training data, it can be easily applied to designing Gabor wavelets for general pattern recognition task.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130974083","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}
Dynamic deformation of target is a prominent problem in image-based tracking. Most existing particle filtering based tracking algorithms treat deformation parameters of the target as a vector. We have proposed a deformable target tracking algorithm via particle filtering on manifolds, which implements the particle filter with the constraint that the system state lies in a low dimensional manifold: affine Lie group. The sequential Bayesian updating consists in drawing state samples while moving on the manifold geodesics; this provides a smooth prior for the state space change. Then we estimate affine deformation parameters through means on Lie group. Theoretic analysis and experimental evaluations against the tracking algorithm based on particle filtering on vector spaces demonstrate the promise and effectiveness of this algorithm.
{"title":"Tracking Deformable Object via Particle Filtering on Manifolds","authors":"Yunpeng Liu, Guangwei Li, Zelin Shi","doi":"10.1109/CCPR.2008.40","DOIUrl":"https://doi.org/10.1109/CCPR.2008.40","url":null,"abstract":"Dynamic deformation of target is a prominent problem in image-based tracking. Most existing particle filtering based tracking algorithms treat deformation parameters of the target as a vector. We have proposed a deformable target tracking algorithm via particle filtering on manifolds, which implements the particle filter with the constraint that the system state lies in a low dimensional manifold: affine Lie group. The sequential Bayesian updating consists in drawing state samples while moving on the manifold geodesics; this provides a smooth prior for the state space change. Then we estimate affine deformation parameters through means on Lie group. Theoretic analysis and experimental evaluations against the tracking algorithm based on particle filtering on vector spaces demonstrate the promise and effectiveness of this algorithm.","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"541 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123084325","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}
A new feature extraction method for high dimensional data using least squares support vector regression (LSSVR) is presented. Firstly, the expressions of optimal projection vectors are derived into the same form as that in the LSSVR algorithm by specially extending the feature of training samples. So the optimal projection vectors could be obtained by LSSVR. Then, using the kernel tricks, the data are mapped from the original input space to a high dimensional feature, and nonlinear feature extraction is here realized from linear version. Finally, it is proved that 1) the method presented has the same result as principal component analysis (PCA). 2) This method is more suitable for the higher dimensional input space compared. 3) The nonlinear feature extraction of the method is equivalent to kernel principal component analysis (KPCA).
{"title":"A Novel Feature Extraction Method and Its Relationships with PCA and KPCA","authors":"Deihui Wu","doi":"10.1109/CCPR.2008.19","DOIUrl":"https://doi.org/10.1109/CCPR.2008.19","url":null,"abstract":"A new feature extraction method for high dimensional data using least squares support vector regression (LSSVR) is presented. Firstly, the expressions of optimal projection vectors are derived into the same form as that in the LSSVR algorithm by specially extending the feature of training samples. So the optimal projection vectors could be obtained by LSSVR. Then, using the kernel tricks, the data are mapped from the original input space to a high dimensional feature, and nonlinear feature extraction is here realized from linear version. Finally, it is proved that 1) the method presented has the same result as principal component analysis (PCA). 2) This method is more suitable for the higher dimensional input space compared. 3) The nonlinear feature extraction of the method is equivalent to kernel principal component analysis (KPCA).","PeriodicalId":292956,"journal":{"name":"2008 Chinese Conference on Pattern Recognition","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127146017","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：