{"title":"NMF vs ICA for face recognition","authors":"Menaka Rajapakse, Lnnce Wyse, Heng Mui Keng","doi":"10.1109/ISPA.2003.1296348","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296348","url":null,"abstract":"This paper deals with the application of spatially localized, nonoverlapping features for face recognition. The analysis is carried out by using the features generated from two closely related techniques known as independent component analysis (ICA) and nonnegative matrix factorization (NMF). A set of statistically independent basis vectors with sparse features is derived from ICA. Likewise, NMF is used to yield sparse representation of localized features to represent distributed parts over a human face. Similarities between reconstructed faces of test images and a set of synthesised face representations from the basis vectors derived from an image database using the two techniques are measured. The strengths and weaknesses of each method in the context of face recognition are discussed.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130811182","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}
{"title":"Design of linear-phase two-channel FIR filter banks with rational sampling factors","authors":"R. Bregović, T. Saramaki","doi":"10.1109/ISPA.2003.1296376","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296376","url":null,"abstract":"The design of two-channel filter banks with rational sampling factors is considered based on the use of linear-phase finite-impulse response (FIR) filters whose stop-band performances are minimized in the least squares sense. It is shown that a proper selection of the filter orders enables one to synthesize a perfect reconstruction filter bank. The optimization of the overall filter bank is carried out by properly modifying the algorithm of Dutta and Vidyasagar. The modified algorithms enables one to optimize both the perfect-reconstruction and the near perfect-reconstruction filter banks. Several examples are included illustrating the properties of the resulting optimized filter banks as well as the trade-off between the stopband attenuation and the amplitude and the aliasing errors of the optimized filter bank.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131016089","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}
{"title":"Fast rate-distortion estimation and optimization for wavelet video coding","authors":"E. S. Jackson, R. Peplow","doi":"10.1109/ISPA.2003.1296900","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296900","url":null,"abstract":"The motion compensated- and difference- residual frames of a typical video sequence displays strong energy localization in the spatial domain. SPIHT expends many bits describing the location of the areas of energy concentration, and at low rates, these bits, which do not contribute to a PSNR increase, are wasteful. The scheme to be described avoids this wastage by performing local rate-distortion estimation and spatial bit assignment to achieve global PSNR optimality. The key to the algorithm is an extremely fast and accurate R-D estimation scheme, which permits this algorithm to operate in real-time.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125349667","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}
{"title":"Bayesian supervised segmentation of objects in natural images using low-level information","authors":"J. Boldys, J. Boldys","doi":"10.1109/ISPA.2003.1296451","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296451","url":null,"abstract":"Detection of particular meaningful objects in images is of great importance in many fields, including image retrieval or image quality analysis. In this contribution, eleven frequently observed objects (areas) in natural images are learned and detected. The presented algorithm is based on region merging and Bayesian decision theory. The main goal is not perfect recognition, since for that purpose it is necessary to use higher-level knowledge about the image content. Merging of segments proceeds only up to a reliable point, not to merge different categories. Unique merging criteria combine the values of probabilities attached to the segments for all the most likely categories, color and texture features and information about edges. Results are demonstrated on a few images and discussed.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126558897","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}
{"title":"Generation of two-dimensional background noise by using computer","authors":"M. Tsudagawa","doi":"10.1109/ISPA.2003.1296866","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296866","url":null,"abstract":"The purpose of this paper is to present the approach to get the two-dimensional background noise with the desired statistical nature by using a computer. This approach gives the picture of the desired two-dimensional background noise by applying the specific filter to the white Gaussian random noise. The filter function is obtained by the mathematical analysis. Since the center value of the filter function becomes infinite, the filter function is modified partially to be able to use actually. Several digital images with the desired statistical nature are generated by this approach. And it is verified that the generated images have the desired statistical parameters.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124986870","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}
{"title":"The direct integer factorization approach to the Crochiere and Rabiner multistage FIR designs for multirate systems","authors":"D. Huang","doi":"10.1109/ISPA.2003.1296453","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296453","url":null,"abstract":"The parameters of designing a multistage, multirate narrow-band FIR system are the total storage requirements (TSR) for the coefficients or the number of multiplications per second (MPS). Crochiere and Rabiner had developed an approximated optimization problem for the choice of parameters. The minimum solution can be found by any suitable multidimensional optimization technique. However, the dimension (number of stages) can only be determined after a full investment of the minimum solutions, and then a heuristic procedure should also be employed to choose the integer factors. An equivalent constrained optimization problem is solved by the direct factorization of positive integer. The proposed algorithm provides a direct factorization to find the true minimum solution of the constrained integer optimization problem. With this approach, the true integer solution of the constrained is obtained. A typical design curve is present by the proposed approach.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122688547","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}
I. Stephanakis, G. Anastassopoulos, A. Karayiannakis, C. Simopoulos
{"title":"Enhancement of medical images using a fuzzy model for segment dependent local equalization","authors":"I. Stephanakis, G. Anastassopoulos, A. Karayiannakis, C. Simopoulos","doi":"10.1109/ISPA.2003.1296420","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296420","url":null,"abstract":"A novel model for fuzzy equalization of medical images is proposed. The method requires a preprocessing step which accomplishes a fuzzy segmentation of the image into N segments with overlapping fuzzy boundaries. Histogram equalization is performed according to individual equalization functions derived from each segment of the image. A fuzzy membership function is associated with each segment. Enhancement results compare well against histogram equalization techniques that derive the equalization transform function from the global histogram of the image.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"207 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121854886","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}
{"title":"Fractal image coding for images using B-splines","authors":"T. Asahi","doi":"10.1109/ISPA.2003.1296895","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296895","url":null,"abstract":"The standard fractal image coding algorithm consists of, given any discrete image specified by an array of pixels, determining a mapping of contractive transformations where the decoded image is visually close to the original one. In this work, the original formulation for discrete images to the space of continuous images based on B-splines is extended. By selecting an appropriate B-spline order n, visually improved images may he obtained, compared to the discrete image approach. This lies on the fact that the blockiness of the edges of the range blocks is diminished due to the entanglement of the B-spline base and the continuity of spline functions. Moreover, the PSNR value of the reconstructed image may be increased at no additional information of the fractal coding and slightly more calculation realized on the space of B-spline coefficients. The results were obtained using a real digital image.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132735617","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}
{"title":"Shift-invariant multichannel blind restoration","authors":"F. Šroubek, J. Flusser","doi":"10.1109/ISPA.2003.1296918","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296918","url":null,"abstract":"Existing multichannel blind restoration techniques are prone to noise, assume perfect spatial alignment of channels and a correct estimation of blur size. We develop an alternating minimization scheme based on maximum a posteriori probability estimation with a priori distribution of blurs derived from the multichannel framework and a priori distribution of original images defined by the total variation semi-norm. This stochastic approach enables us to recover the blurs and the original image from channels severely corrupted by noise. We observe that the exact knowledge of the blur size is not necessary and we show that translation misregistration up to a certain extent can be automatically removed in the restoration process.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"185 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133762210","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}
{"title":"A new 2D adaptive nonlinear filter based on the Lyapunov stability theory","authors":"S. Sakrani, M. Sayadi, F. Fnaiech","doi":"10.1109/ISPA.2003.1296470","DOIUrl":"https://doi.org/10.1109/ISPA.2003.1296470","url":null,"abstract":"In this paper, a new 2D adaptive nonlinear filter is proposed. Its stability is guaranteed using Lyapunov stability theory. This algorithm profits of the matrix structure of the window (mask), used to define the 2D signal as a nonlinear model of exponential form. This nonlinear exponential model may be easily expanded in a Taylor series leading a higher order polynomial filter. Using the Lyapunov stability theory, it is shown that the new algorithm is independent of the stochastic character of the input signal. Simulation results highlight the efficiency of the new algorithm.","PeriodicalId":218932,"journal":{"name":"3rd International Symposium on Image and Signal Processing and Analysis, 2003. ISPA 2003. Proceedings of the","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114399997","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}