Browsing by Author "Pao, HT"

Now showing 1 - 11 of 11

A Self-growing Probabilistic decision-based neural network for anchor/speaker identification
(2003)
Chen, YH
;
Tseng, CL
;
Fu, HC
;
Pao, HT
;
資訊工程學系
In this paper, we propose a new clustering algorithm for a mixture Gaussian based neural network, called Self-growing Probabilistic decision-based neural networks (SPDNN). The proposed Self-growing cluster learning (SGCL) algorithm is able to find the natural number of prototypes based on a self-growing validity measure, Bayesian Information Criterion (BIC). The learning process starts with a single prototype randomly initialized in the feature space and grows adaptively during the learning process until most appropriate number of prototypes are found. We have conduct numerical and real world experiments to demostrate the effectiveness of the SGCL algorithm. In the results of using SGCL to trainin the SPDNN for anchor/speaker identification, we have observed noticeable improvement among various model-based or vector quantization-based classification schemes.
A self-growing probabilistic decision-based neural network with automatic data clustering
(2004-10-01)
Tseng, CL
;
Chen, YH
;
Xu, YY
;
Pao, HT
;
Fu, HC
In this paper, we propose a new clustering algorithm for a mixture of Gaussian-based neural network and self-growing probabilistic decision-based neural networks (SPDNN). The proposed self-growing cluster learning (SGCL) algorithm is able to find the natural number of prototypes based on a self-growing validity measure, Bayesian information criterion (BIC). The learning process starts from a single prototype randomly initialized in the feature space and grows adaptively during the learning process until most appropriate number of prototypes are found. We have conducted numerical and real-world experiments to demonstrate the effectiveness of the SGCL algorithm. In the results of using SGCL to train the SPDNN for data clustering and speaker identification problems, we have observed a noticeable improvement among various model-based or vector quantization-based classification schemes. (C) 2004 Elsevier B.V. All rights reserved.
A WWW-based intelligent multimedia information query and retrieve system
(2000)
Xu, YY
;
Fu, HC
;
Pao, HT
;
資訊工程學系
;
Department of Computer Science
In this paper, we propose and implement a World Wide Web based video query system: the intelligent Multimedia Information Processing System (IMIPS). The IMIPS first collects videos available over the WWW, analyzes these videos to extract keyframes, and indexes these keyframes at a local database. By using the proposed Multiple-Instance learning algorithm, a user can select a few keyframes to train a desired image query model. When a query model is selected, the proposed IMIPS will retrieve marched images from the local database, and display them as a small icon at the user interface. Then a user can click on the interested icon to retrieve and to play the video from the original web site. According to the implemented prototype system, a user can find one's desired videos on the WWW with only a few (2-3) times of retraining to achieve 70% hit rate on 50 retrieved videos among 5 different query models.
Adaptive watermarking using relationships between wavelet coefficients
(2005)
Chen, YH
;
Su, JM
;
Fu, HC
;
Huang, HC
;
Pao, HT
In this paper, we propose an approach that hides watermarks in relationships between wavelet coefficients. This approach would minimize the perceptual distortion of watermarked images, measured with PSNR or just noticeable distortion (JND). Therefore, the strength of watermarks can be enlarged to increase the robustness of the watermarks, while keeping the quality of watermarked images visually accepted. Experimental results show that the watermark is still detectable after common image processing operations such as JPEG compression, Gaussian filtering and sharpening.
Automated information. mining on multimedia TV news archives
(2005)
Lai, PS
;
Cheng, SS
;
Sun, SY
;
Huang, TY
;
Su, JM
This paper addresses an integrated information mining techniques for multimedia TV-news archive. The utilizes techniques from the fields of acoustic, image, and video analysis, for information retrieval on news story title, newsman and scene identification. The goal is to construct a compact yet meaningful abstraction of broadcast news video, allowing users to browse through large amounts of data in a non-linear fashion with flexibility and efficiency. By using acoustic analysis, the system can classify video into news versus commercials, with 90% accuracy on a data set of 400 hours TV-news recorded off the air from July 2003 to August of 2004. By applying speaker identification and/or image detection techniques, each news stories can be segmented with an accuracy of 96%. On screen captions or subtitles are recognized by OCR techniques to produce the text title of each news stories. The extracted title words can be used to link or to navigate more related News contents on the WWW. In cooperation with facial and scene analysis and recognition techniques, OCR results can provide users with multimodality query for specific news stories. Some experimental results are presented and discussed for the system reliability and performance evaluation and comparison.
Comparison of TSCS regression and neural network models for panel data forecasting: debt policy
(2006-04-01)
Pao, HT
;
Chih, YY
;
管理科學系
;
Department of Management Science
Empirical studies of variations in debt ratios across firms have analyzed important determinants of capital structure using statistical models. Researchers, however, rarely employ nonlinear models to examine the determinants and make little effort to identify a superior prediction model among competing ones. This paper reviews the time-series cross-sectional (TSCS) regression and the predictive abilities of neural network (NN) utilizing panel data concerning debt ratio of high-tech industries in Taiwan. We built models with these two methods using the same set of measurements as determinants of debt ratio and compared the forecasting performance of five models, namely, three TSCS regression models and two NN models. Models built with neural network obtained the lowest mean square error and mean absolute error. These results reveal that the relationships between debt ratio and determinants are nonlinear and that NNs are more competent in modeling and forecasting the test panel data. We conclude that NN models can be used to solve panel data analysis and forecasting problems.
Divide-and-conquer learning and modular perceptron networks
(2001-03-01)
Fu, HC
;
Lee, YP
;
Chiang, CC
;
Pao, HT
;
資訊工程學系
A novel modular perceptron network (MPN) and divide-and-conquer learning (DCL) schemes for the design of modular neural networks are proposed, When a training process in a multilayer perceptron fails into a local minimum or stalls in a flat region, the proposed DCL scheme is applied to divide the current training data region (e.g,, a hard to be learned training set) into two easier (hopely) to be learned regions. The learning process continues when a self-growing perceptron network and its initial weight estimation are constructed for one of the newly partitioned regions, Another partitioned region will resume the training process on the original perceptron network. Data region partitioning, weight estimating and learning are iteratively repeated until all the training data are completely learned by the MPN We have evaluated and compared the proposed MPN with several representative neural networks on the two-spirals problem and real-world dataset, The MPN achieves better weight learning performance by requiring much less data presentations (99.01% similar to 87.86% lesser) during the network training phases, and better generalization performance (4.0% better), and less processing time (2.0% similar to 81.3% lesser) during the retrieving phase, On learning the real-world data, the MPN's show less overfitting compared to single MLP. In addition, due to its self-growing and fast local learning characteristics, the modular network (MPN) can easily adapt to on-line and/or incremental Learning requirements for a rapid changing environment.
Face detection and eye localization by neural network based color segmentation
(2000)
Fu, HC
;
Lai, PS
;
Lou, RS
;
Pao, HT
;
資訊工程學系
This paper presents a neural network based scheme for human face detection and eye localization in color images under non-constrained scene. A Self-growing Probabilistic Decision-based Neural Network (SPDNN) is used to learn the conditional distribution for each color classes. Pixels of a color image are first classified into facial or non-facial regions, then pixels in the facial region are followed by eye region segmentation. The class of each pixel is determined by using the conditional distribution of the chrominance components of pixels belonging to each class. The paper demonstrates a successful application of SPDNN to face detection and eye localization on a database of 755 images from 151 persons. Regarding the performance, experimental results are elaborated in Section 3. As to the processing speed, the face detection and eye localization processes consume approximately 560 ms on a Pentium-II personal computer.
Forecasting electricity market prices: A neural network based approach
(2004)
Xu, YY
;
Hsieh, R
;
Lu, Y
;
Shen, YC
;
Chuang, SC
This paper presents a neural network approach to forecast the Phelix Base (PB) electricity market prices for European Energy Exchange (EEX). Up to now there has been little scientific work on forecasting the price development on the electricity markets. In this study, the Phelix Base moving average (PBMA), the moving difference (PBMD), and multilayer feedforward neural networks (MLNN) are used to predict various period for 7, 14, 21, 28, 63, 91, 182, and 273 days ahead of electric prices. The experimental results of forecasting by MLNNs and linear methods (autoregressive error model) are compared and discussed. The MLNNs outperform from 11.4% to 64.6% superior to the traditional linear regression method. It seems that the proposed MLNN can be very useful in predicting the electricity market prices of EEX.
User adaptive handwriting recognition by self-growing probabilistic decision-based neural networks
(2000-11-01)
Fu, HC
;
Chang, HY
;
Xu, YY
;
Pao, HT
;
資訊工程學系
It is generally agreed that, for a given handwriting recognition task, a user dependent system usually outperforms a user independent system, as long as a sufficient amount of training data is available. When the amount of user training data is limited, however, such a performance gain is not guaranteed. One way to improve the performance is to make use of existing knowledge, contained in a rich multiuser data base, so that a minimum amount of training data is sufficient to initialize a model for the new user, We mainly address the user adaption issues for a handwriting recognition system, Based on self-growing probabilistic decision-based neural networks (SPDNNs), user adaptation of the parameters of SPDNN is formulated as incremental reinforced and antireinforced learning procedures, which are easily integrated into the batched training procedures of the SPDNN, In this study, me developed 1) an SPDNN based handwriting recognition system; 2) a two-stage recognition structure; and 3) a three-phase training methodology for a) a global coarse classifier (stage 1); b) a user independent hand written character recognizer (stage 2); and c) a user adaptation module on a personal computer. With training and testing on a 600-word commonly used Chinese character set, the recognition results indicate that the user adaptation module significantly improved the recognition accuracy The average recognition rate increased from 44.2% to 82.4% in five adapting cycles, and the performance could finally increase up to 90.2% in ten adapting cycles.
Using statistical and neural network methods to explore the relationship between systematic risk and firm's long term investment activities
(2000)
Wang, K
;
Pao, HT
;
Fu, HC
;
管理科學系
;
Department of Management Science
This paper presents a neural network method to explore the relationship between the systematic risk and the long term investing activities for Taiwan's companies in the fiber industry and in the electronics industry. In general, diversification from long-term investment may reduce firm's systematic risk, but the empirical results in some literatures are controversial. For years regression methods have been used to analyze the possible impact of long-term investment activities on systematic risk. Since their relationship may not be linear we thus propose a neural network based sensitivity analysis for the possible non-linear relationships. We adapt 5 years data (1994-1998) from the TEJ financial database to conduct the proposed analysis. The results show that the systematic risk is reduced with investment activities for the fiber industry. But for the electronics industry, the systematic risk is higher as firms increase the long-term investment ratio. The difference is even more significant when we only consider the companies with higher portion of long term investment in assets. This diverse effect between industries may be one of the reasons why the empirical results are inconsistent. Our study can clarify the controversial between the systematic risk and the long-term investment activities, and offer some possible explanations.