Sj Yeom

Elimination of redundant information for Web data mining

These days, billions of Web pages are created with HTML or other markup languages. They only have a few uniform structures and contain various authoring styles compared to traditional text-based documents. However, users usually focus on a particular section of the page that presents the most relevant information to their interest. Therefore, Web documents classification needs to group and filter the pages based on their contents and relevant information. Many researches on Web mining report on mining Web structure and extracting information from Web contents. However, they have focused on detecting tables that convey specific data, not the tables that are used as a mechanism for structuring the layout of Web pages. Case modeling of tables can be constructed based on structure abstraction. Furthermore, Ripple Down Rules (RDR) is used to implement knowledge organization and construction, because it supports a simple rule maintenance based on case and local validation.

Security mechanism for medical image information on PACS using invisible watermark

The emerging digital revolution is impacting all areas of society, including medical areas. The era of digital hospital is coming and its success will depend on the rapid adaptation of emerging technology. Picture Archive and Communication System(PACS) could give the necessary competitive edge for medical institutions. Pacs digitizes captured images using the technique in Computed Radiography, stores them in massive storage devices, and transmits the patients medical information(or images) to networked clinics and hospital wards, anywhere anytime. However, to be useful, the issues of possible fabrication and damages to images, and authentication have to be addressed. The watermark technique could provide some solution to these problems. This paper describles an improved secure medical image system addressing issues on PACS, accurate patient data, and privacy. Further, the proof of concept is demonstrated by experiment.

A Study on Warning/Detection Degree of Warranty Claims Data Using Neural Network Learning

Warranty service is getting important since it is an agreement between manufacturers and consumers. An issue is to find out a lower level of agreement from the perspective of manufacturers and consumers. Thus, it is very important to determine early warning/detection degree of defected parts through warranty claims data. However, there are qualitative factors more than quantitative ones in the determination. The study thus provides a part-significance knowledge extraction method based on analytic hierarchy process analysis which is appropriate to analyze those qualitative factors as well as a process to extract a list of defected parts using neural network learning.

Object modeling for mapping XML document represented in XML-GDM to UML class diagram

XML has been popular as a means of sharing and distributing data due to its flexible and open architecture. XML-GL, a visual and intuitive query language for XML document, is easily used to search structures of XML documents and share information, since it represents the semantics of query and the structure of found documents visually. UML is used as a tool to analyze and design an object oriented system via defined notation and various diagrams. In this paper, we will propose a new object modeling method to map XML documents based on XML-GDM (a data model of XML-GL) to UML class diagrams. Thus, XML documents can be converted and stored and managed into object oriented data by an intuitive method. Applying the object oriented search method will improve the effectiveness in search of XML documents.

Factors influencing undergraduate students’ acceptance of a haptic interface for learning gross anatomy

Purpose This study aims to evaluate factors influencing undergraduate students’ acceptance of a computer-aided learning resource using the Phantom Omni haptic stylus to enable rotation, touch and kinaesthetic feedback and display of names of three-dimensional (3D) human anatomical structures on a visual display. Design/methodology/approach The software was developed using the software development life cycle, and was tested by students enrolled in various bachelor degrees at three stages of development within the technology acceptance model, action research and design research methodology frameworks, using mixed methods of quantitative and qualitative analysis. Findings The learning system was generally well-accepted, with usefulness (72 ± 18, mean ± standard deviation, 0-100 visual analogue scale) rated higher (p < 0.001) than ease of use (57 ± 22). Ease of use ratings declined across the three versions as modules were added and complexity increased. Students with prior experience with 3D interfaces had higher intention to use the system, and scored higher on identification of anatomical structures. Students with greater kinaesthetic learning preferences tended to rate the system higher. Haptic feedback was considered the best aspect of the system, but students wanted higher spatial resolution and lower response times. Originality/value Previous research relating to haptic devices in medical and health sciences has largely focused on advanced trainees learning surgical or procedural skills. The present research suggests that incorporating haptic feedback into virtual anatomical models may provide useful multisensory information in learning anatomy at the undergraduate level.

Design of a Robust Watermarking Algorithm against the Geometric Distortion for Medical Image Security

This study suggests using a selected geometric invariant point through the whole processing procedure of an image and inserting and extracting based on the invariant point so that it will be robust in a geometric transformation attack. The introduced algorithm here is based on a watershed splittling method in order to make medical images strong against RST transformation and other processing. It also helps to maintain the watermark in images that are compressed and stored for a period of time.

Traffic Sign Recognition Using Color Information and Error Back Propagation Algorithm

In this thesis, the color information is used to extract the traffic sign territory, and for recognizing the extracted image, it proposes the traffic sign recognition system that applies the error back propagation algorithm. The proposed method analyzes the color of traffic sign to extract and recognize the possible territory of traffic sign. The method of extracting the possible territory is to use the characteristics of YUV, YIQ, and CMYK color space from the RGB color space. Morphology uses the geometric characteristics of traffic sign to make the image segmentation. The recognition of traffic signs can be recognized by using the error back propagation algorithm. As a result of the experiment, the proposed system has proven its outstanding capability in extraction and recognition of candidate territory without the influence of differences in lighting and input image in various sizes.

Effective delivery of virtual class on parallel media stream server

A virtual class delivers most of its content through multimedia learning objects. To support such multimedia learning objects, its multimedia data service system should have a capacity to serve the growing number of clients and new data. A streaming server transfers multimedia files to clients from a repository of files in real time. The server must guarantee concurrent and uninterrupted delivery of each video stream requested by clients. To provide efficient services, many stream servers adopt multi-processors, sufficient memory, and RAID or SAN in their systems. In this paper, we propose a Linux-based parallel media streaming server. This system uses a unique striping policy to distribute multimedia files into the parallel storage nodes. If a service request occurs, each storage node transmits striped files concurrently to the client. Its performance is better than the existing single media streaming service.

ECG-based biometric human identification based on backpropagation neural network

Biometric human identifications are expansively reshaping security applications in the emerging sophisticated era of smart devices. To inflate the level of security and privacy demands, human physiological signal based human identification and authentication systems are getting tremendous attention. This study focuses on producing feasible amount of segmented signals from a source signal for training dataset, and integrating 2-layer framework backpropagation neural network to handle the great amount of classes for identification without hesitation. The results suggest that the proposed method surpasses the recent technique with the similar architecture, and possesses more advantages in terms of computational complexity and high performance compared with the previously reported study.

Exploring the Use of Big Data Analytics for Improving Support to Students in Higher Education

In the past two decades, with the globalisation of education, there has been a continuous increase in the diversity of students in Higher Education. This diversity form a basis for a culturally rich environment, although, the cultural and language differences and the diversity in teaching and learning styles also bring challenges. From a university’s perspective, providing the maximum support to overcome these challenges and achieving maximised student engagement would be in its best interest. Recent advances in Big Data and increase in electronically available education data can help in achieving these aims. This paper reports the findings of a preliminary study which applies Big Data analysis methods to analyse education data gathered from learning management systems. The aims was to understand ways to improve student engagement and reduce student dropout. This paper documents the experience gained in this early exploration and preliminary analysis, and thereby provides background knowledge for reporting of data from the formal data collection stage which will be conducted at a later stage of research.