{"title":"Dynamically pruning output weights in an expanding multilayer perceptron neural network","authors":"H. Amin, K. M. Curtis, B.R. Hayes Gill","doi":"10.1109/ICDSP.1997.628530","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628530","url":null,"abstract":"The network size for a multilayer perceptron neural network is often chosen arbitrarily for different applications, and the optimum size of the network is determined by a long process of trial and error. This paper presents a backpropagation algorithm. For a multilayer perceptron (MLP) neural network, that dynamically determines the optimum number of hidden nodes and applies a new pruning technique on output weights. A 29% reduction in the total number of output weights was observed for a handwritten character recognition problem using the new pruning algorithm.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133553474","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 fast skeleton algorithm on block represented binary images","authors":"Iraklis M. Spiliotis, Basil G. Mertzios","doi":"10.1109/ICDSP.1997.628440","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628440","url":null,"abstract":"This paper describes a binary image representation scheme, which is called image block representation and presents a new skeletonization algorithm, which is implemented on block represented binary images. The main purpose of the image block representation is to provide an efficient binary image representation that permits the execution of operations on image areas instead of image points. The skeletonization algorithm operates in four subiterations: each subiteration deletes the north, the south, the west and the east boundary points, respectively. Due to the substitution of the boundary points by the block's boundary points the relevant operations are performed fast, while preserving the end points and the object connectivity.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133251779","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 symmetry-based approach to facial features extraction","authors":"F. D. De Natale, D. Giusto, F. Maccioni","doi":"10.1109/ICDSP.1997.628390","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628390","url":null,"abstract":"Face recognition technology (FRT) is gaining increasing interest in a large number of application fields. The article deals with a new approach to this problem, that exploits symmetry characteristics of the human visage to extract a set of significant features for the detection task. The developed methods give a very good percentage of success in the identification of facial elements, with the advantage of fast computation, that makes the proposed approach suitable for real-time applications, like personal identification and communications.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115127842","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":"Sinusoidal signal detection using the minimum description length and the predictive stochastic complexity","authors":"S. Valaee, B. Champagne, P. Kabal","doi":"10.1109/ICDSP.1997.628538","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628538","url":null,"abstract":"Two techniques based on the minimum description length (MDL) and the predictive stochastic complexity (PSC) are proposed for sinusoidal signal detection. The MDL and PSC criteria are the codelength of the observation and the model. The proposed techniques decompose the observation vector into its components in the signal and noise subspaces. The noise component is encoded for several model orders. The best model is selected by minimizing the codelength.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116352715","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":"FWPT filtering properties","authors":"L. Carracciuolo, L. D’Amore, A. Murli","doi":"10.1109/ICDSP.1997.628572","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628572","url":null,"abstract":"Discusses the possible use of the fast wavelet transform in order to develop effective preconditioners as well as efficient regularization schemes for image restoration.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123419041","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 momentary Fourier transformation derived from recursive matrix transformations","authors":"S. Albrecht, Ian G. Cumming, J. Dudás","doi":"10.1109/ICDSP.1997.628089","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628089","url":null,"abstract":"The momentary Fourier transform (MFT) computes the DFT of a discrete-time sequence for every new sample of the sequence. It has an efficient recursive form, and an alternate derivation is given using matrix transformations. A recursive form of the inverse MFT is also given, which is particularly efficient as it involves no multiplications.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127290673","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":"Constrained adaptive multichannel L-filters","authors":"C. Kotropouloz, M. Gabram, I. Pitas","doi":"10.1109/ICDSP.1997.628063","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628063","url":null,"abstract":"Adaptive multichannel L-filters based on marginal ordering when structural constraints such as location invariance or unbiasedness are imposed on the filter coefficients are proposed. Two novel adaptive algorithms are derived by using the Frost (1972) algorithm for minimizing the mean squared error subject to constraints in LMS and in LMS-Newton algorithms. It is demonstrated by experiments that the Frost-LMS algorithm has a faster convergence rate than the Frost LMS-Newton algorithm but it yields a higher steady-state MSE.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129148701","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}
N. Ikonomatakis, K. Plataniotis, M. Zervakis, A. Venetsanopoulos
{"title":"Region growing and region merging image segmentation","authors":"N. Ikonomatakis, K. Plataniotis, M. Zervakis, A. Venetsanopoulos","doi":"10.1109/ICDSP.1997.628077","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628077","url":null,"abstract":"Image segmentation is an important first task of any image analysis process. This paper presents a seeded region growing and merging algorithm that was created to segment grey scale and colour images. The approach starts with a set of seed pixels and from these grows regions by appending to each seed pixel those neighbouring pixels that satisfy a certain predicate. Small regions of far away values were merged to neighbouring regions while regions of similar value were also merged. Homogeneity functions are introduced for both grey scale and colour images.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132059406","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":"Enhancement of the Horn and Schunck optic flow algorithm by means of median filters","authors":"F. Bartolini, A. Piva","doi":"10.1109/ICDSP.1997.628385","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628385","url":null,"abstract":"For optic flow computation the Horn and Schunck (1980) algorithm is one of the most used and powerful. Some drawbacks are, nevertheless, intrinsically related to the approach that it used. A smoothness constraint is imposed to the flow field in order to produce a solution. This constraint causes the estimated field to be incorrect where motion discontinuities are present. The weight of this smoothness constraint is reduced by using median filters. Motion boundaries are better preserved by the modified algorithm. Results are presented and possible developments are proposed.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123159884","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":"Efficient direction estimation of uncorrelated sources using polynomial rooting","authors":"B. Goransson, B. Ottersten","doi":"10.1109/ICDSP.1997.628517","DOIUrl":"https://doi.org/10.1109/ICDSP.1997.628517","url":null,"abstract":"This paper presents a novel technique for estimating directions of arrival of uncorrelated narrowband sources. It is also shown that, if a uniform linear array is assumed, the estimation can be performed by a computationally attractive rooting technique. It is further shown that the asymptotic covariance of the estimation error attains the proper Cramer-Rao lower bound. The theoretical results presented are compared to computer simulations.","PeriodicalId":220163,"journal":{"name":"Proceedings of 13th International Conference on Digital Signal Processing","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1997-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126218043","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}