Pub Date : 1996-10-14DOI: 10.1109/ICSIGP.1996.567290
A. Cichocki, W. Kasprzak, S. Amari
In this paper an adaptive approach to the cancellation of additive, convolutional noise from many-source mixtures with simultaneous blind source separation is proposed. Associated neural network learning algorithms are developed on the basis of the decorrelation principle and energy minimization of the output signals. The reference noise is transformed into convolutional noise by employing an adaptive FIR filter in each channel. Several models of NN learning processes are considered. In the basic approach the noisy signals are separated simultaneously with additive noise cancellation. The simplified model employs separate learning steps for noise cancellation and source separation. Multi-layer neural networks improve the quality of the results. The results of comparative tests of the proposed methods are provided.
{"title":"Adaptive approach to blind source separation with cancellation of additive and convolutional noise","authors":"A. Cichocki, W. Kasprzak, S. Amari","doi":"10.1109/ICSIGP.1996.567290","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.567290","url":null,"abstract":"In this paper an adaptive approach to the cancellation of additive, convolutional noise from many-source mixtures with simultaneous blind source separation is proposed. Associated neural network learning algorithms are developed on the basis of the decorrelation principle and energy minimization of the output signals. The reference noise is transformed into convolutional noise by employing an adaptive FIR filter in each channel. Several models of NN learning processes are considered. In the basic approach the noisy signals are separated simultaneously with additive noise cancellation. The simplified model employs separate learning steps for noise cancellation and source separation. Multi-layer neural networks improve the quality of the results. The results of comparative tests of the proposed methods are provided.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"204 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123050853","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.566280
Sun Xinding, Wang Yan, Xu Zhenming, X. Changsheng
In the view of problems by processing the lateral cephalometric films in the automatic cephalometric analysis system, a noise-removing Kalman filter in spatial non-homogeneous images is proposed. The noise is effectively reduced and the edges in the image are preserved by adding information about the edges to the input of the filter. This approach can avoid the computational burden caused by the model parameter identification.
{"title":"Research on spatial non-homogenous Kalman filter and its application in cephalometric image","authors":"Sun Xinding, Wang Yan, Xu Zhenming, X. Changsheng","doi":"10.1109/ICSIGP.1996.566280","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.566280","url":null,"abstract":"In the view of problems by processing the lateral cephalometric films in the automatic cephalometric analysis system, a noise-removing Kalman filter in spatial non-homogeneous images is proposed. The noise is effectively reduced and the edges in the image are preserved by adding information about the edges to the input of the filter. This approach can avoid the computational burden caused by the model parameter identification.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125086119","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.566232
H. Hetzheim
This paper is concerned with the identification of corresponding points in curves related together, as time series, or in epipolar lines of stereo images. The properties of the corresponding points are defined by the near neighbourhood but also further points which are related to the corresponding point. Filtering with special non-linearities is applied to suppress non-relevant information and to find interrelations. For this the curves are represented by non-linear stochastic differential equations. The properties of the curves are obtained by the estimation of the expectation values of these stochastic equations and represented by different fuzzy measures. The corresponding points are determined by fusion of information obtained from different properties related to a special corresponding point with the help of fuzzy integrals. Using filtering and fuzzy integration of properties, feature and intensity based methods are combined.
{"title":"Fusion of intensity and feature based analysis for matching of corresponding points","authors":"H. Hetzheim","doi":"10.1109/ICSIGP.1996.566232","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.566232","url":null,"abstract":"This paper is concerned with the identification of corresponding points in curves related together, as time series, or in epipolar lines of stereo images. The properties of the corresponding points are defined by the near neighbourhood but also further points which are related to the corresponding point. Filtering with special non-linearities is applied to suppress non-relevant information and to find interrelations. For this the curves are represented by non-linear stochastic differential equations. The properties of the curves are obtained by the estimation of the expectation values of these stochastic equations and represented by different fuzzy measures. The corresponding points are determined by fusion of information obtained from different properties related to a special corresponding point with the help of fuzzy integrals. Using filtering and fuzzy integration of properties, feature and intensity based methods are combined.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125096611","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.567262
E. Grall-Maes, P. Beauseroy
This paper presents a method to remove cross-components produced by the discrete Wigner-Ville distribution (WVD). The procedure consists of considering the WVD as an image and assigning each pixel to either an auto-component or a cross-component according to a correlation coefficient. This coefficient measures the correlation between the time-frequency representations of the local signal content and of a chirp Gaussian signal. While the auto-components yield large coefficient values, the cross-components yield small values due to their oscillating structure. The representation is obtained from the WVD by keeping pixels whose value is positive and the correlation coefficient larger than a threshold. It has the advantage of being positive, characterized by a high concentration and no distortion of the auto-components. This method provides good performance with a large class of signals.
{"title":"Elimination of cross-components of the discrete Wigner-Ville distribution via a correlation method","authors":"E. Grall-Maes, P. Beauseroy","doi":"10.1109/ICSIGP.1996.567262","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.567262","url":null,"abstract":"This paper presents a method to remove cross-components produced by the discrete Wigner-Ville distribution (WVD). The procedure consists of considering the WVD as an image and assigning each pixel to either an auto-component or a cross-component according to a correlation coefficient. This coefficient measures the correlation between the time-frequency representations of the local signal content and of a chirp Gaussian signal. While the auto-components yield large coefficient values, the cross-components yield small values due to their oscillating structure. The representation is obtained from the WVD by keeping pixels whose value is positive and the correlation coefficient larger than a threshold. It has the advantage of being positive, characterized by a high concentration and no distortion of the auto-components. This method provides good performance with a large class of signals.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125138182","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.566327
Gong Jian, Liang Liyuan, Chen Weinan
Two-dimensional (2D) thresholding behaves well in segmenting images of low signal-to-noise ratio. But the computational complexity of the conventional 2D entropic algorithm is bounded by O(L/sup 4/). Firstly, a fast recursive 2D entropic thresholding algorithm is proposed. By rewriting the formula for calculation of the entropy in a recurrence form, a great deal of calculation is saved. Analysis shows that the computational complexity of 2D entropic thresholding is reduced to O(L/sup 2/). The fast recursive algorithm is also used successfully in the 2D Otsu (1979) method. Experimental results show that the processing time of each image is reduced from more than 2 h to less than 10 sec. The required memory space is also greatly reduced.
{"title":"A fast recursive algorithm for two-dimensional thresholding","authors":"Gong Jian, Liang Liyuan, Chen Weinan","doi":"10.1109/ICSIGP.1996.566327","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.566327","url":null,"abstract":"Two-dimensional (2D) thresholding behaves well in segmenting images of low signal-to-noise ratio. But the computational complexity of the conventional 2D entropic algorithm is bounded by O(L/sup 4/). Firstly, a fast recursive 2D entropic thresholding algorithm is proposed. By rewriting the formula for calculation of the entropy in a recurrence form, a great deal of calculation is saved. Analysis shows that the computational complexity of 2D entropic thresholding is reduced to O(L/sup 2/). The fast recursive algorithm is also used successfully in the 2D Otsu (1979) method. Experimental results show that the processing time of each image is reduced from more than 2 h to less than 10 sec. The required memory space is also greatly reduced.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123542826","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.566340
E. Trouvé, J. Nicolas, H. Maître
Least squares methods provide a smart global solution to the two dimensional phase unwrapping problem, but results obtained on synthetic aperture radar (SAR) interferograms are usually undervalued because of several kinds of perturbations. In this paper, we analyze the bias due to noise corruption of wrapped phase differences. To avoid it and to prevent the solution from the influence of inconsistent areas, we propose to use an estimation of the local frequency as phase gradient and the associated measure of confidence as a mask. With these robust input data, weighted least squares unwrapping becomes an automatic method useful for SAR interferometric applications.
{"title":"Noise immune least squares phase unwrapping based on two dimensional frequency estimation","authors":"E. Trouvé, J. Nicolas, H. Maître","doi":"10.1109/ICSIGP.1996.566340","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.566340","url":null,"abstract":"Least squares methods provide a smart global solution to the two dimensional phase unwrapping problem, but results obtained on synthetic aperture radar (SAR) interferograms are usually undervalued because of several kinds of perturbations. In this paper, we analyze the bias due to noise corruption of wrapped phase differences. To avoid it and to prevent the solution from the influence of inconsistent areas, we propose to use an estimation of the local frequency as phase gradient and the associated measure of confidence as a mask. With these robust input data, weighted least squares unwrapping becomes an automatic method useful for SAR interferometric applications.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125295775","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.567311
Mao Yongcai, B. Zheng
Harmonic retrieval is one of the most frequently encountered problems in practice, and constitutes a significant part of statistical signal processing research. The concern here is retrieval of multiple tone harmonics observed in white Gaussian multiplicative and white complex Gaussian additive noise. Computable Cramer-Rao bound (CRB) expressions are derived on the frequency and phase estimates as well as on the sample mean of the multiplicative noise processes. Numerical studies support the obtained results.
{"title":"Cramer-Rao bounds for harmonics in Gaussian multiplicative and complex Gaussian additive noise","authors":"Mao Yongcai, B. Zheng","doi":"10.1109/ICSIGP.1996.567311","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.567311","url":null,"abstract":"Harmonic retrieval is one of the most frequently encountered problems in practice, and constitutes a significant part of statistical signal processing research. The concern here is retrieval of multiple tone harmonics observed in white Gaussian multiplicative and white complex Gaussian additive noise. Computable Cramer-Rao bound (CRB) expressions are derived on the frequency and phase estimates as well as on the sample mean of the multiplicative noise processes. Numerical studies support the obtained results.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126746371","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.571147
D. Ryu, C. Kim, S.W. Kim, T. Choi, J.C. Kim
This paper presents a new search region prediction method using the neural networks vector quantization (VQ) in the motion estimation. A major advantage of formulating VQ as neural networks is that the large number of adaptive training algorithm that are used for neural networks can be applied to VQ. The proposed method reduces the computation because of the smaller number of search points than conventional methods, and reduces the bits required to represent motion vectors. The results of computer simulation show that the proposed method provides better PSNR than other block matching algorithms.
{"title":"Motion search region prediction using neural network vector quantization","authors":"D. Ryu, C. Kim, S.W. Kim, T. Choi, J.C. Kim","doi":"10.1109/ICSIGP.1996.571147","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.571147","url":null,"abstract":"This paper presents a new search region prediction method using the neural networks vector quantization (VQ) in the motion estimation. A major advantage of formulating VQ as neural networks is that the large number of adaptive training algorithm that are used for neural networks can be applied to VQ. The proposed method reduces the computation because of the smaller number of search points than conventional methods, and reduces the bits required to represent motion vectors. The results of computer simulation show that the proposed method provides better PSNR than other block matching algorithms.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"223 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115563963","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.566333
J. Zhu, H. Sun, Z. Chai
In this paper, a 3-channel 8-bit analog HDTV (high definition television) image sampling and recording system is described in detail. It can grab images up to 1024/spl times/1024 pixels from a motion picture sequence (in the case of 2:1 interlaced scanning; 25 frames per second) with low error bit rate. Finally, the test result is presented.
{"title":"A wide range video A/D converter and processor for non-broadcasting digital HDTV system","authors":"J. Zhu, H. Sun, Z. Chai","doi":"10.1109/ICSIGP.1996.566333","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.566333","url":null,"abstract":"In this paper, a 3-channel 8-bit analog HDTV (high definition television) image sampling and recording system is described in detail. It can grab images up to 1024/spl times/1024 pixels from a motion picture sequence (in the case of 2:1 interlaced scanning; 25 frames per second) with low error bit rate. Finally, the test result is presented.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"145 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116106171","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 : 1996-10-14DOI: 10.1109/ICSIGP.1996.567258
Zhang Xudong, Wang Deshang, Peng Ying-ning, Lili Yuan, R. Mark
We propose a new video coding scheme based on four improvements. (1) Half pixel precision is used for the overlapped motion compensation. (2) Region wavelet decomposition is only applied to the significant region in the prediction error image (displacement frame difference, DFD). (3) Zerotree is used for representation of the significant region wavelet coefficients. (4) A quantization model based on the human visual system (HVS) have been used for intraframe coding. Simulation results show that some coding gain has been reached compared to the conventional wavelet coder.
{"title":"Overlapped motion compensated region wavelet coder for very low bit-rates","authors":"Zhang Xudong, Wang Deshang, Peng Ying-ning, Lili Yuan, R. Mark","doi":"10.1109/ICSIGP.1996.567258","DOIUrl":"https://doi.org/10.1109/ICSIGP.1996.567258","url":null,"abstract":"We propose a new video coding scheme based on four improvements. (1) Half pixel precision is used for the overlapped motion compensation. (2) Region wavelet decomposition is only applied to the significant region in the prediction error image (displacement frame difference, DFD). (3) Zerotree is used for representation of the significant region wavelet coefficients. (4) A quantization model based on the human visual system (HVS) have been used for intraframe coding. Simulation results show that some coding gain has been reached compared to the conventional wavelet coder.","PeriodicalId":385432,"journal":{"name":"Proceedings of Third International Conference on Signal Processing (ICSP'96)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1996-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122604901","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：