Winston Lo

Artifact reduction of JPEG coded images using mean-removed classified vector quantization

Image compression techniques are frequently applied to reduce the network bandwidth and storage space. In the case of higher compression ratios, annoying artifacts may be generated and they degrade the perceptual quality of compressed images. This paper modified mean-removed classified vector quantization (MRCVQ) to reduce the artifacts of JPEG coded images. This algorithm consists of four phases: mean removal, encoding, decoding, and mean restoration. The mean removal phase removes the mean values of compressed image blocks. The encoding procedure needs a codebook for the encoder, which transforms a mean-removed compressed image to a set of codeword-indices. The decoding phase requires a different codebook for the decoder, which enhances a mean-removed compressed image from a set of codeword-indices. Finally, the mean values are restored in the mean restoration phase. The experimental results show that the proposed approach can remove effectively the artifacts caused by high compression and improve the perceptual quality significantly. Compared to the existing methods, our approach usually has the much better performance in terms of computing time, storage space and PSNR.

A new multi-search engine for querying data through an Internet search service on CORBA

Abstract Search engines are important but generally far from ideal tools of the World Wide Web (WWW). Many researchers therefore prefer to use meta-brokers to construct multi-search engines (MSE). However, these have no uniform programming interfaces, which makes tying them with other search engines difficult. Moreover, for an application that needs a search service capability, querying them is difficult. To reduce that difficulty, we propose in this paper an Internet search service (ISS) based on common object request broker architecture (CORBA) that follows the style of common object service specification (COSS). We design a multi-search engine based on ISS, which we term Octopus. For a system developer, because of its uniformity of interface, Octopus easily ties with any search engine. Equally, for an application programmer, the ISS offers a clear interface for application programs to search for information or mine data from the Internet. We demonstrate our approach to designing multi-search engines through ISS by tying two search engine agents, Yahoo and AltaVista, with Octopus and show how CORBA clients query them. Programmers may use this interface to construct their search engine agents or query a search engine agent in their applications.

Building CORBA objects with DOS software applications

CORBA is becoming the most important middleware that supports the object-oriented and client/server paradigm in distributed computing systems. However, the application systems based on CORBA are still scarce. One main reason is that only a few CORBA object services have been developed. To develop a new CORBA application, a programmer should make an effort to design a program with a CORBA interface from scratch. In our previous work (Liang et al., 1997), a re-engineering approach was proposed to convert RPC-based programs to CORBA objects, which successfully speeds up the development of CORBA applications. However, the source code is required in this approach. In many cases, software designers may not get hold of the source code. It will be significant to adapt existing PC application software in binary code mode to the object services under CORBA. Our study is addressing this problem. A graphical parking lot management system, which integrates dBase III and AutoCAD V2.6i under DOS has been implemented to demonstrate the feasibility of our approach.

On the Accelerated Convergence of Genetic Algorithm Using GPU Parallel Operations

The genetic algorithm plays a very important role in many areas of applications. In this research, the authors propose to accelerate the evolution speed of the genetic algorithm by parallel computing, and optimize parallel genetic algorithms by methods such as the island model. The authors find that when the amount of population increases, the genetic algorithm tends to converge more rapidly into the global optimal solution; however, it also consumes greater amount of computation resources. To solve this problem, the authors take advantage of the many cores of GPUs to enhance computation efficiency and develop a parallel genetic algorithm for GPUs. Different from the usual genetic algorithm that uses one thread for computation of each chromosome, the parallel genetic algorithm using GPUs evokes large amount of threads simultaneously and allows the population to scale greatly. The large amount of the next generation population of chromosomes can be divided by a block method; and after independently operating in each block for a few generation, selection and crossover operations of chromosomes can be performed among blocks to greatly accelerate the speed to find the global optimal solution. Also, the travelling salesman problem TSP is used as the benchmark for performance comparison of the GPU and CPU; however, the authors did not perform algebraic optimization for TSP.

An Information Store and Retrieval Facility on CORBA

In this paper, we propose an Information Storage and Retrieval Facility based on the CORBA. This facility provides information search, retrieval and storage service. To define these services, we referred the Z39.50 and defined a set of interfaces by CORBA IDL. In addition, other information storage respect services, such as information update, information update notify, and information version control etc., and more other services, such as initial service, close service, and search command forward and distribution etc., were provided in this facility. Based on this facility, programmer can add the capability of query Z39.50 server into applications and also can integrate various data sources and provide an uniform interface. Finally, we also implement an experimental library management system to verify our design.

An Agent-Based Personalized Search on a Multi-Search Engine Based on Internet Search Service

The Internet Search Service (ISS) was proposed to support an uniform interface for searching on the World Wide Web. Based on this service, a multi-search engine named Octopus had been built. In order to provide more services, such as personal functionality, we provide personalized search for users. In this paper, the policies of personalized search are described. In addition, in order to keep the advantages of the ISS, a personal information-filtering agent is added into the Octopus instead of modifying the architecture or interface of ISS. The feedback mechanism is in coopcratiori with the filfering nrechanism to achieve the functionally of personalized search in a search engine.

Design and implementation of multi-threaded object request broker

The distributed object oriented computing model is the next logical step to develop distributed applications. In recent years, several object models have been proposed, such as COM/DCOM, CORBA, and JAVA Bean etc. In CORBA, which was announced by OMG, object request broker is a software bus to connect applications and object components. In addition, multi threaded programming is a well known technique to improve the performance of applications. In a CORBA environment, clients can invoke the remote objects that are shared. If those objects are single threaded it will affect system performance in large distributed applications. We describe in detail the design and implementation of multi threaded object request broker based on CORBA. Our ORB was implemented atop Windows NT and underlying TCP protocol. Finally, we compare our systems performance with IONAs Orbix, which is a well known commercial product, in both one-way and two-way request.

Artifact reduction of compressed color images using modified mean-removed classified vector quantization

Abstract In this paper, a new artifact reduction algorithm for compressed color images using MMRCVQ is proposed. The algorithm extends and modifies vector quantization (VQ) for discovering the relationships between the uncompressed color images and their deblocked compressed versions by classifying the deblocked compressed blocks into several categories using information from their neighboring blocks. The discovered relationships are stored in two codebooks and are used to recover the missing information of compressed color images. To increase the availability of codewords and reduce the memory needed for storing codewords, mean‐removed vectors are used to generate codebooks. The experimental results show that the proposed approach can remove, effectively, the artifacts caused by high compression and improve perceptual quality significantly. Compared to existing methods, the proposed approach usually uses much less computing time to recover a compressed color image and has much better image quality.

Image restoration of JPEG coded images using mean removed classified vector quantization

Image compression techniques are frequently applied for reducing the requirement of network bandwidth and storage space. To remove the artifacts and improve the quality of compressed images, a mean-removed classified vector quantization (MRCVQ) algorithm is proposed. The algorithm extends and modifies vector quantization (VQ) to discover the relationships between the uncompressed color images and their deblocked compressed versions by classifying the deblocked compressed blocks into several categories using information from their neighboring blocks. The discovered relationships are stored in two codebooks and used to recover the missing information of compressed color images. The experimental results show that the proposed algorithm can remove the artifacts and improve the quality of compressed color images effectively.

Two-level Web agent for limited accessibility

Many problems frustrate WWW users, who suffer limited accessibility to WWW resources, and so the authors propose a two-level agent architecture to try to solve them. The agent consists of two parts: the local agent interacts directly with end-users, while the remote agent interacts primarily with the public Internet. Together they learn user preferences, and help users find and filter information more efficiently, economically, and enjoyably. The paper introduces the concept of the two-level agent, and investigates practical issues involved in further design.