Showing papers by "Jeng-Shyang Pan published in 2009"

Enhanced artificial bee colony optimization

Abstract: An enhanced Artificial Bee Colony (ABC) optimization algorithm, which is called the Interactive Artificial Bee Colony (IABC) optimization, for numerical optimiza- tion problems, is proposed in this paper. The onlooker bee is designed to move straightly to the picked coordinate indicated by the employed bee and evaluates the fitness values near it in the original Artificial Bee Colony algorithm in order to reduce the computa- tional complexity. Hence, the exploration capacity of the ABC is constrained in a zone. Based on the framework of the ABC, the IABC introduces the concept of universal grav- itation into the consideration of the affection between employed bees and the onlooker bees. By assigning different values of the control parameter, the universal gravitation should be involved for the IABC when there are various quantities of employed bees and the single onlooker bee. Therefore, the exploration ability is redeemed about on average in the IABC. Five benchmark functions are simulated in the experiments in order to com- pare the accuracy/quality of the IABC, the ABC and the PSO. The experimental results manifest the superiority in accuracy of the proposed IABC to other methods.

Forecasting enrollments using automatic clustering techniques and fuzzy logical relationships

Abstract: In recent years, some researchers focused on the research topic of using fuzzy time series to handle forecasting problems In this paper, we present a new method to forecast enrollments based on automatic clustering techniques and fuzzy logical relationships First, we present an automatic clustering algorithm for clustering historical enrollments into intervals of different lengths Then, each obtained interval will be divided into p sub-intervals, where p>=1 Based on the new obtained intervals and fuzzy logical relationships, we present a new method for forecasting the enrollments of the University of Alabama The proposed method gets a higher average forecasting accuracy rate than the existing methods

Weighted Fuzzy Interpolative Reasoning Based on Weighted Increment Transformation and Weighted Ratio Transformation Techniques

Abstract: In this paper, we present a new weighted fuzzy interpolative reasoning method for sparse fuzzy rule-based systems. The proposed method uses weighted increment transformation and weighted ratio transformation techniques to handle weighted fuzzy interpolative reasoning in sparse fuzzy rule-based systems. It allows each variable that appears in the antecedent parts of fuzzy rules to associate with a weight between zero and one. Moreover, we also propose an algorithm that automatically tunes the optimal weights of the antecedent variables appearing in the antecedent parts of fuzzy rules. We also apply the proposed weighted fuzzy interpolative reasoning method to handle the truck backer-upper control problem. The proposed weighted fuzzy interpolative reasoning method performs better than the ones obtained by the traditional fuzzy inference system (2000), Huang and Shen's method (2008), and Chen and Ko's method (2008). The proposed method provides us with a useful way to deal with weighted fuzzy interpolative reasoning in sparse fuzzy rule-based systems.

Kernel optimization-based discriminant analysis for face recognition

Abstract: The selection of kernel function and its parameter influences the performance of kernel learning machine. The difference geometry structure of the empirical feature space is achieved under the different kernel and its parameters. The traditional changing only the kernel parameters method will not change the data distribution in the empirical feature space, which is not feasible to improve the performance of kernel learning. This paper applies kernel optimization to enhance the performance of kernel discriminant analysis and proposes a so-called Kernel Optimization-based Discriminant Analysis (KODA) for face recognition. The procedure of KODA consisted of two steps: optimizing kernel and projecting. KODA automatically adjusts the parameters of kernel according to the input samples and performance on feature extraction is improved for face recognition. Simulations on Yale and ORL face databases are demonstrated the feasibility of enhancing KDA with kernel optimization.

Fourth International Conference on Innovative, Computing, Information and Control

Abstract: The following topics are dealt with: innovation management; information network; e-learning; image recognition; information retrieval; intelligent system; artificial intelligence; data mining; and motor control

Innovations in Digital Watermarking Techniques

Abstract: Information security and copyright protection are more important today than before. Digital watermarking is one of the widely used techniques used in the world in the area of information security. This book introduces a number of digital watermarking techniques and is divided into four parts. The first part introduces the importance of watermarking techniques and intelligent technology. The second part includes a number of watermarking techniques. The third part includes the hybrid watermarking techniques and the final part presents conclusions. This book is directed to students, professors, researchers and application engineers who are interested in the area of information security.

Face recognition using Gabor-based complete Kernel Fisher Discriminant analysis with fractional power polynomial models

Abstract: This paper presents a novel face recognition method by integrating the Gabor wavelet representation of face images and the enhanced powerful discriminator, complete Kernel Fisher Discriminant (CKFD) with fractional power polynomial (FPP) models. The novelty of this paper comes from (1) Gabor wavelet, is employed to extract desirable facial features characterized by spatial frequency, spatial locality and orientation selectivity to cope with the variations in illumination and facial expressions, which improves the recognition performance; (2) a recently proposed powerful discriminator, namely CKFD, which enhances its discriminating ability using two kinds of discriminant information (i.e., regular and irregular information), is employed to classify the Gabor features; (3) the FPP models, are employed to CKFD analysis to enhance the discriminating ability. Comparing with existing principal component analysis, linear discriminant analysis, kernel principal component analysis, KFD and CKFD methods, the proposed method gives the superior results with the ORL, Yale and UMIST face databases.

A Particle Swarm Optimization with Feasibility-Based Rules for Mixed-Variable Optimization Problems

Abstract: A Particle Swarm Optimization algorithm with feasibility-based rules (FRPSO) is proposed in this paper to solve mixed-variable optimization problems. An approach to handle various kinds of variables is discussed. Constraint handling is based on simple feasibility-based rules, not needing addinional penalty parameters and not guaranteeing to be in the feasible region at all times. Two real-world mixed-varible optimization benchmark problems are presented to evaluate the performance of the FRPSO algorithm, and it is found to be highly competitive compared to other existing stochastic algorithms.

Information Hiding and Applications

Abstract: Genetic Watermarking for Copyright Protection.- Dual-Plane Correlation-Based Video Watermarking for Immunity to Rotation, Scale, Translation, and Random Distortion.- Restoring Objects for Digital Inpainting.- A Secure Data Embedding Scheme Using Gray-Code Computation and SMVQ Encoding.- Robust Image Watermarking Based on Scale-Space Feature Points.- Intelligent Perceptual Shaping in Digital Watermarking.- Semi-fragile Image Authentication Method for Robust to JPEG, JPEG2000 Compressed and Scaled Images.- Genetic-Based Fingerprinting for Multicast Multimedia.- Lossless Data Hiding for Halftone Images.- Information Hiding by Digital Watermarking.

Digital Watermarking Techniques

Abstract: In this chapter, some basic concepts about digital watermarking are described in as general terms as possible. An introduction is given first. The classification of watermarking, the functions provided by watermarking, and the benchmarks and evaluating measurements for watermarking systems are then introduced in the following sections.

Parity-invariability-based reversible watermarking

Abstract: A reversible integer transform, defined on an image pixel pair, is proposed. The transform can carry 1-bit watermark while retaining the parity of the sum of the pairs. Also, the transform makes one transformed pixel not related to the watermark bit which can be reused to construct a new pair with its next pixel. So the embedding rate can approach 1 bit per pixel for a single embedding process. Meanwhile, the decision on whether this new pair can be used for reversible embedding or not is greatly simplified. In addition, the application of the transform is able to largely increase the compression ratio of the location map, and finally increase the payload. A series of experiments has been conducted to verify the effectiveness of the proposed approach.

An Improved Particle Swarm Optimization with Feasibility-Based Rules for Constrained Optimization Problems

Abstract: This paper presents an improved particle swarm optimization (IPSO) to solve constrained optimization problems, which handles constraints based on certain feasibility-based rules. A turbulence operator is incorporated into IPSO algorithm to overcome the premature convergence. At the same time, a set called FPS is proposed to save those P best locating in the feasible region. Different from the standard PSO, g best in IPSO is chosen from the FPS instead of the swarm. Furthermore, the mutation operation is applied to the P best with the maximal constraint violation value in the swarm, which can guide particles to close the feasible region quickly. The performance of IPSO algorithm is tested on a well-known benchmark suite and the experimental results show that the proposed approach is highly competitive, effective and efficient.

A New Vector Particle Swarm Optimization for Constrained Optimization Problems

Abstract: A new vector Particle Swarm Optimization (NVPSO) is proposed to solve constrained optimization problems in this paper. In NVPSO, we introduce a shrinkage coefficient to ensure that all dimensions of a particle are within lower and upper bounds, and a new function to determine whether the particle is within the feasible region. One-dimensional search optimization methods are selected in NVPSO algorithm to produce a new position which is guaranteed to be in the feasible region for the particle which escapes from the feasible region. The whole process is dealt as vector mode. The experimental results show that the principle of NVPSO this paper proposed is simple, relatively effective and efficient.

CSO and PSO to solve optimal contract capacity for high tension customers

Abstract: The aim of this research is to study the optimal demand contract decision for the Taiwanese industries through CSO and PSO algorithm Yet CSO algorithm can be proposed to select optimal contract capacity and drop the basic electricity cost Results indicated that the CSO algorithm is highly helpful to Taiwanese industries on the optimal demand contract decision Also the CSO is superior to PSO in the fast convergence and better performance to find the global best solution

Genetic Watermarking for Copyright Protection

Abstract: Applications for robust watermarking is one of the major branches in digital rights management (DRM) systems. Based on existing experiences to assess how good one robust watermarking is, it is generally agreed that three parameters or requirements, including the quality of watermarked contents, the survivability of extracted watermark after deliberate or unintentional attacks, and the number of bits embedded, need to be considered. However, performances relating to these three parameters conflict with each other, and the trade off must be searched for. In this chapter, we take these requirements into consideration, and we can find the optimized combination among the three parameters. With the aid of genetic algorithm, we design an applicable system that would obtain the good quality, acceptable survivability, and reasonable capacity after watermarking. Simulation results present the effectiveness in practical implementation and possible application of the proposed algorithm.

An Improved Particle Swarm Optimization with Feasibility-Based Rules for Mixed-Variable Optimization Problems

Abstract: This paper presents an improved particle swarm optimization algorithm with feasibility- based rules (FRIPSO) to solve mixed-variable constrained optimization problems. Different kinds of variables are dealt in different ways in FRIPSO algorithm. Constraint handling is based on simple feasibility-based rules without the use of a penalty function which is frequently cumbersome to parameterize, nor need it to guarantee the particles be in the feasible region at all time which turn out to cost much time sometimes. In order to improve the convergence speed of FRIPSO with the iteration growing and to find global optimum, the standard PSO is used to find a better position for the best history position of the swarm on the condition that the discrete value are same with those of Gbest in each iteration. Two practical benchmark mixed-variable optimization problems are tested by our FRIPSO algorithm to demonstrate the effectiveness and robustness of the proposed approach.

Lossless data hiding based on prediction-error adjustment

Abstract: A novel lossless data hiding scheme based on a combination of prediction and the prediction-error adjustment (PEA) is presented in this paper. For one pixel, its four surrounding neighboring pixels are used to predict it and 1-bit watermark information is embedded into the prediction-error. In traditional approaches, for the purpose of controlling embedding distortion, only pixels with small prediction-errors are used for embedding. However, when the threshold is small, it is difficult to efficiently compress the location map which is used to identify embedding locations. Thus, PEA is introduced to make large prediction-error available for embedding while causing low embedding distortions, and accordingly, the location map can be compressed well. As a result, the hiding capacity is largely increased. A series of experiments are conducted to verify the effectiveness and advantages of the proposed approach.

Scale-Space Feature Based Image Watermarking in Contourlet Domain

Abstract: In scale-space feature based watermarking schemes, the watermark is usually embedded in spatial domain so that watermark robustness is not satisfactory. In this paper, a novel image watermarking scheme is presented by combining scale-space feature based watermark synchronization and nonsubsampled Contourlet transform (NSCT) based watermark embedding. Watermark synchronization is achieved based on the local circular regions, which can be generated using the scale-invariant feature transform (SIFT). In the encoder, the watermark is embedded into the NSCT coefficients in a content-based and rotation-invariant manner by odd-even quantization. In the decoder, the watermark can be extracted directly from the local regions using the proposed coefficient property detector (CPD). Simulation results and comparisons show that the proposed scheme can efficiently resist signal processing attacks, geometric attacks as well as some combined attacks.

The Application and Study of Graph Cut in Motion Segmentation

Abstract: How to segment moving objects accurately and rapidly is one of the most important problems in computer vision. In this paper, we discuss two methods studied previously for segmenting moving objects based on graph cut. One method employs graph cut to segment objects automatically, and the other uses the graph cut and the C-V model (simplified Mumford-Shah model) to segment objects. As well, we present a fast segmentation algorithm based on graph cut. The proposed algorithm is trying to reduce the number of the nodes in constructing network graph. By mapping the invariable pixels of the difference image into one or several nodes and mapping the variable pixels into other nodes, the number of the nodes and the edges is decreased and the speed of graph cut is increased. In the end of the paper, we compare the characteristics of three methods. Experimental results show that we can get effective and fast results in motion segmentation by using graph cut technology.

Robust Image Watermarking Based on Scale-Space Feature Points

Abstract: Digital watermarking techniques have been explored extensively since its first appearance in the 1990s. However, watermark robustness to geometric attacks is still an open problem. The past decade has witnessed a significant improvement in the understanding of geometric attacks and how watermarks can survive such attacks. In this chapter, we will introduce a set of image watermarking schemes which can resist both geometric attacks and traditional signal processing attacks simultaneously. These schemes follow a uniform framework, which is based on the detection of scale-space feature points. We call it the scale-space feature point based watermarking, SSFW for short. Scale-space feature points have been developed recently for pattern recognition applications. This kind of feature points are commonly invariant to rotation, scaling and translation (RST), therefore they naturally fit into the framework of geometrically robust image watermarking. Scale-space feature points are typically detected from the scale space of the image. As a result, we will first introduce the scale space theory and how the feature points can be extracted. The basic principles on how the scale-space feature points can be adapted for watermark synchronization are then discussed in detail. Subsequently, we will present several content-based watermark embedding and extraction methods which can be directly implemented based on the synchronization scheme. A detailed watermarking scheme which combines scale-invariant feature transform (SIFT) and Zernike moments is then presented for further understanding of SSFW. Watermarking schemes based on the SSFW framework have the following advantages: (a) Good invisibility. The Peak Signal to Noise Ratio (PSNR) value is typically higher than 40dB. (b) Good robustness. These schemes can resist both signal processing attacks and geometric attacks, such as JPEG compression, image filtering, added noise, RST attacks, locally cropping as well as some combined attacks.

RBF Neural Network Based on Fuzzy Evolution Kalman Filtering and Application in Mine Safety Monitoring

Abstract: Fuzzy information fusion methods are adopted widely to resolve the complicated nonlinear problems in recent years. This paper proposes a fusion learning algorithm of radial basis function (RBF) neural network based on fuzzy evolution Kalman filtering. By using this proposed method, monitoring data are extracted and optimized in mine safety monitoring, and Matlab simulation results are analyzed. The results show that this method has feasibility and rapid learning efficiency, which can improve precision and reliability in mine monitoring systems.

A new method for constrained optimization problems to produce initial values

Abstract: To solve the hard problem that it is difficult to produce initial values satisfied with the constrained conditions in constrained optimization problems, according to the good ability of particle swarm optimization in finding good values, the paper presents a new method for producing stochastic values satisfied with constrained conditions using particle swarm optimization, which can be used in the all kinds of algorithms to produce initial values. The examples show that the algorithm this paper presents can get good stochastic values satisfied with constrained conditions.

A load management method to reduce the power energy cost for glass industry

Abstract: The aim of this research is to study the power cost reduction of glass manufacturing process for industrial customers and to relieve the crisis of electricity shortage during summer peak time period through load management (LM) This research can also provide decision makers and managers with useful LM strategies to use as guidance The results indicate that local glass industry is highly willing to participate in the development of LM strategies As the running time of a ball mill is adjusted to the preferential rates of time of use (TOU), industry customers can save a certain amount of electricity cost

Watermarking with Visual Cryptography and Gain-Shape VQ

Abstract: Employing the concept of visual cryptography to design a watermarking scheme is introduced in this chapter. A modified visual cryptography is applied to split the genuine watermark into two shadow watermarks. The gain-shape vector quantisation (GSVQ) procedure is then performed to encode the cover image. Afterwards, the shadow watermarks and the VQ indices obtained are processed to generate two user-keys.

PSO algorithm applications in optimal demand decision

Abstract: The aim of this research is to select the optimal demand contract for the TOU rate customer through PSO algorithm and to drop the electricity cost for the Taiwanese industry. PSO is an alternative population-based evolutionary computation technique, also it is has more fast convergence than GA to select optimal contract capacity for TOU rate customer. Results indicated that the PSO has been highly willing to Taiwanese industry in the optimal demand contract decision. Also this research aims at measuring the benefit on TOU rate and to provide decision-makers and managers with useful load management strategies as reference.

Spatial-Based Watermarking Schemes and Pixel Selection

Abstract: Spatial domain based watermarking is one of the fundamental techniques at the beginning of digital watermarking. Generally speaking, spatial domain based watermarking schemes possess the advantages, such as easy implementation and low complexity, than other domain based watermarking schemes. But also, they posses the shortages like weak robustness and non-practical usage than other domain based watermarking schemes.

Image Segmentation Based on GC-CV

Abstract: In this article, an integrated method, named GC-CV, was developed and applied to image segmentation. The proposed method combines graph cut method and the simplified Mumford-Shah model (C-V model), and takes the advantages of both. In this paper, the proposed GC-CV method is put to test in its three different operational modes. The first mode is the segmentation of binary images using GC-CV directly. The second mode is the segmentation of multi-region images by recursive GC-CV. The last one is segmentation of color images and gray images by combining GC-CV with EM algorithm, and using YCbCr color space for color image segmentation. The feasibility and effectiveness of the proposed GC-CV method is verified by a serious of experiments. Experimental results show that the speed of segmentation can be greatly improved and the number of iterations can be considerably reduced with GC-CV in comparison with C-V model. Experimental results reveal that GC-CV can be a promising approach to image segmentation.

The triple-watermarking algorithm with multiple description coding over lossy communication networks

Abstract: Watermarking is one useful solution for digital rights management (DRM) systems In this paper, in addition to following conventional objectives to conquer against intentional or unintentional attacks, we not only watch the survivability of embedded watermark, but also focus on retaining the watermarked image quality with the aid of multiple description coding (MDC) MDC is an algorithm for error resilient coding, suitable for transmitting compressed data over multiple channels In this paper, we propose an innovative algorithm on vector quantization (VQ) based image watermarking, which is suitable for error-resilient transmission over noisy channels By incorporating watermarking with multiple description coding, the scheme we proposed for embedding three watermarks can effectively overcome channel impairments while retaining the capability for copyright and ownership protection With the promising simulation results presented, we can demonstrate the utility and practicability of our algorithm