• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.1-23
• /
• 2018

From the 21st century, various high-quality services have come up with the growth of the internet or 'Information and Communication Technologies'. Especially, the scale of E-commerce industry in which Amazon and E-bay are standing out is exploding in a large way. As E-commerce grows, Customers could get what they want to buy easily while comparing various products because more products have been registered at online shopping malls. However, a problem has arisen with the growth of E-commerce. As too many products have been registered, it has become difficult for customers to search what they really need in the flood of products. When customers search for desired products with a generalized keyword, too many products have come out as a result. On the contrary, few products have been searched if customers type in details of products because concrete product-attributes have been registered rarely. In this situation, recognizing texts in images automatically with a machine can be a solution. Because bulk of product details are written in catalogs as image format, most of product information are not searched with text inputs in the current text-based searching system. It means if information in images can be converted to text format, customers can search products with product-details, which make them shop more conveniently. There are various existing OCR(Optical Character Recognition) programs which can recognize texts in images. But existing OCR programs are hard to be applied to catalog because they have problems in recognizing texts in certain circumstances, like texts are not big enough or fonts are not consistent. Therefore, this research suggests the way to recognize keywords in catalog with the Deep Learning algorithm which is state of the art in image-recognition area from 2010s. Single Shot Multibox Detector(SSD), which is a credited model for object-detection performance, can be used with structures re-designed to take into account the difference of text from object. But there is an issue that SSD model needs a lot of labeled-train data to be trained, because of the characteristic of deep learning algorithms, that it should be trained by supervised-learning. To collect data, we can try labelling location and classification information to texts in catalog manually. But if data are collected manually, many problems would come up. Some keywords would be missed because human can make mistakes while labelling train data. And it becomes too time-consuming to collect train data considering the scale of data needed or costly if a lot of workers are hired to shorten the time. Furthermore, if some specific keywords are needed to be trained, searching images that have the words would be difficult, as well. To solve the data issue, this research developed a program which create train data automatically. This program can make images which have various keywords and pictures like catalog and save location-information of keywords at the same time. With this program, not only data can be collected efficiently, but also the performance of SSD model becomes better. The SSD model recorded 81.99% of recognition rate with 20,000 data created by the program. Moreover, this research had an efficiency test of SSD model according to data differences to analyze what feature of data exert influence upon the performance of recognizing texts in images. As a result, it is figured out that the number of labeled keywords, the addition of overlapped keyword label, the existence of keywords that is not labeled, the spaces among keywords and the differences of background images are related to the performance of SSD model. This test can lead performance improvement of SSD model or other text-recognizing machine based on deep learning algorithm with high-quality data. SSD model which is re-designed to recognize texts in images and the program developed for creating train data are expected to contribute to improvement of searching system in E-commerce. Suppliers can put less time to register keywords for products and customers can search products with product-details which is written on the catalog.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.12
• /
• pp.55-64
• /
• 2013

In this paper, the HILS (Hardware In-the-Loop Simulation) system to analyze and to verify the performance of the targeting pod is addressed. The main functions of the targeting pod is acquiring and tracking targets to guide a LGB (Laser Guided Bomb) to the targets. For the analysis of targeting pod, the real time simulate images generation of IR and daylight cameras, sever control technology, and the analysis of laser transfer characteristics are necessary. For the real time image generation and the laser transfer characteristics analysis, off-the-shelf SDK(Software Development Kit) OKTAL-SE is used. For the servo controller, well-proven mechanism in the previous program is applied to increase servo control accuracy. To analyze the performance of a targeting pod in a realistic environment, 1553B, ARINK818 interface and etc. which are actually implemented in real combat aircrafts are applied in the system. By using the developed HILS system, the performance of currently operating targeting pods in real combat aircrafts can be analyzed and predicted. Additionally, the relationship between overall system performance and each module performance can be analyzed, the currently developed HILS system is expected to be a very useful tool to generate system development requirements of targeting pods and to reduce any possible future development risks.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.2C
• /
• pp.182-192
• /
• 2010

This paper proposes the hardware IP design of JBIG2 encoder. In order to facilitate the next generation FAX after the standardization of JBIG2, major modules of JBIG2 encoder are designed and implemented, such as symbol extraction module, Huffman coder, MMR coder, and MQ coder. ImpulseC Codeveloper and Xilinx ISE/EDK program are used for the synthesis of VHDL code. To minimize the memory usage, 128 lines of input image are processed succesively instead of total image. The synthesized IPs are downloaded to Virtex-4 FX60 FPGA on ML410 development board. The four synthesized IPs utilize 36.7% of total slice of FPGA. Using Active-HDL tool, the generated IPs were verified showing normal operation. Compared with the software operation using microblaze cpu on ML410 board, the synthesized IPs are better in operation time. The improvement ratio of operation time between the synthesized IP and software is 17 times in case of symbol extraction IP, and 10 times in Huffman coder IP. MMR coder IP shows 6 times faster and MQ coder IP shows 2.2 times faster than software only operation. The synthesized H/W IP and S/W module cooperated to succeed in compressing the CCITT standard document.

• Journal of The Korean Association For Science Education
• /
• v.27 no.1
• /
• pp.93-104
• /
• 2007

The purpose of this study is to investigate brain activity both during the processing of a scientific hypothesis about biological phenomena and mental arithmetic using 3.0T fMRI at the KAIST. For this study, 16 healthy male subjects participated voluntarily. Each subject's functional brain images by performing a scientific hypothesis task and a mental arithmetic task for 684 seconds were measured. After the fMRI measuring, verbal reports were collected to ensure the reliability of brain image data. This data, which were found to be adequate based on the results of analyzing verbal reports, were all included in the statistical analysis. When the data were statistically analyzed using SPM2 software, the scientific hypothesis generating process was found to have independent brain network different from the mental arithmetic process. In the scientific hypothesis process, we can infer that there is the process of encoding semantic derived from the fusiform gyrus through question-situation analysis in the pre-frontal lobe. In the mental arithmetic process, the area combining pre-frontal and parietal lobes plays an important role, and the parietal lobe is considered to be involved in skillfulness. In addition, the scientific hypothesis process was found to be accompanied by scientific emotion. These results enabled the examination of the scientific hypothesis process from the cognitive neuroscience perspective, and may be used as basic materials for developing a learning program for scientific hypothesis generation. In addition, this program can be proposed as a model of scientific brain-based learning.

• Korean Journal of Remote Sensing
• /
• v.36 no.5_2
• /
• pp.893-906
• /
• 2020

This paper reports on the analysis of the location accuracy of a precision image generation system manufactured for CAS 500. The planned launch date of the CAS 500 is 2021, and since it has not yet been launched, the analysis was performed using KOMPSAT-3A satellite images having similar specifications to the CAS 500. In this paper, we have checked the geolocation accuracy of initial sensor model, the model point geolocation accuracy of the precise sensor model, the geolocation accuracy of the precise sensor model using the check point, and the geolocation accuracy of the precise orthoimage using 30 images of the Korean Peninsula. In this study, the target geolocation accuracy is to have an RMSE within 2 pixels when an accurate ground control point is secured. As a result, it was confirmed that the geolocation accuracy of the precision sensor model using the checkpoint was about 1.85 pixels in South Korea and about 2.04 pixels in North Korea, and the geolocation accuracy of the precise orthoimage was about 1.15 m in South Korea and about 3.23 m in North Korea. Overall, it was confirmed that the accuracy of North Korea was low compared to that of South Korea, and this was confirmed to have affected the measured accuracy because the GCP (Ground Control Point) quality of the North Korea images was poor compared to that of South Korea. In addition, it was confirmed that the accuracy of the precision orthoimage was slightly lower than that of precision sensor medel, especially in North Korea. It was judged that this occurred from the error of the DTM (Digital Terrain Model) used for orthogonal correction. In addition to the causes suggested by this paper, additional studies should be conducted on factors that may affect the position accuracy.

• Journal of Intelligence and Information Systems
• /
• v.17 no.1
• /
• pp.139-152
• /
• 2011

Recommender systems have become an important research field since the emergence of the first paper on collaborative filtering in the mid-1990s. In general, recommender systems are defined as the supporting systems which help users to find information, products, or services (such as books, movies, music, digital products, web sites, and TV programs) by aggregating and analyzing suggestions from other users, which mean reviews from various authorities, and user attributes. However, as academic researches on recommender systems have increased significantly over the last ten years, more researches are required to be applicable in the real world situation. Because research field on recommender systems is still wide and less mature than other research fields. Accordingly, the existing articles on recommender systems need to be reviewed toward the next generation of recommender systems. However, it would be not easy to confine the recommender system researches to specific disciplines, considering the nature of the recommender system researches. So, we reviewed all articles on recommender systems from 37 journals which were published from 2001 to 2010. The 37 journals are selected from top 125 journals of the MIS Journal Rankings. Also, the literature search was based on the descriptors "Recommender system", "Recommendation system", "Personalization system", "Collaborative filtering" and "Contents filtering". The full text of each article was reviewed to eliminate the article that was not actually related to recommender systems. Many of articles were excluded because the articles such as Conference papers, master's and doctoral dissertations, textbook, unpublished working papers, non-English publication papers and news were unfit for our research. We classified articles by year of publication, journals, recommendation fields, and data mining techniques. The recommendation fields and data mining techniques of 187 articles are reviewed and classified into eight recommendation fields (book, document, image, movie, music, shopping, TV program, and others) and eight data mining techniques (association rule, clustering, decision tree, k-nearest neighbor, link analysis, neural network, regression, and other heuristic methods). The results represented in this paper have several significant implications. First, based on previous publication rates, the interest in the recommender system related research will grow significantly in the future. Second, 49 articles are related to movie recommendation whereas image and TV program recommendation are identified in only 6 articles. This result has been caused by the easy use of MovieLens data set. So, it is necessary to prepare data set of other fields. Third, recently social network analysis has been used in the various applications. However studies on recommender systems using social network analysis are deficient. Henceforth, we expect that new recommendation approaches using social network analysis will be developed in the recommender systems. So, it will be an interesting and further research area to evaluate the recommendation system researches using social method analysis. This result provides trend of recommender system researches by examining the published literature, and provides practitioners and researchers with insight and future direction on recommender systems. We hope that this research helps anyone who is interested in recommender systems research to gain insight for future research.

• Journal of the Korean Institute of Traditional Landscape Architecture
• /
• v.32 no.2
• /
• pp.148-158
• /
• 2014

This study aims to analyze the Spatial Imagery(Yijing) of the Weeping Bamboo Lodge(Xiaoxiangguan) which is from Chinese Qing dynasty novel "Dream of Red Mansions". The conclusions are as follows. First, the fantasy garden what is described in the novel "Dream of Red Mansions" can be recreated in reality. Second, through the analysis of the spatial imagery, the plants of the Weeping Bamboo Lodge contains a lot of meaning, and mainly through the plants to express meaning. Third, the main garden concept of the Weeping Bamboo Lodge is "Inspired by Nature", the representative space constitution principle is "the art of circuitous" and "view borrowing". Fourth, the concept of traditional garden in the novel "Dream of Red Mansions" and the landscape architecture theory book "Yuan Ye(Art of garden building)" is essentially in agreement. The generation process of garden spatial imagery was showed in this study, and on the basis of this, the garden spatial imagery of the Weeping Bamboo Lodge was analyzed. It is provided the useful information for the future research, and the novel "Dream of Red Mansions" as a important book was determined in the research of traditional garden.

• Journal of Intelligence and Information Systems
• /
• v.17 no.4
• /
• pp.37-58
• /
• 2011

In general, testing DB applications is much more difficult than testing other types of software. The fact that the DB states as much as the input data influence and determine the procedures and results of program testing is one of the decisive reasons for the difficulties. In order to create and maintain proper DB states for testing, it not only takes a lot of time and efforts, but also requires extensive IT expertise and business knowledge. Despite the difficulties, there are not enough research and tools for the needed help. This article reports the result of research on automatic creation and maintenance of DB states for testing DB applications. As its core, this investigation develops an automation tool which collects relevant information from a variety of sources such as log, schema, tables and messages, combines collected information intelligently, and creates pre- and post-Images of database tables proper for application tests. The proposed procedures and tool are expected to be greatly helpful for overcoming inefficiencies and difficulties in not just unit and integration tests but including regression tests. Practically, the tool and procedures proposed in this research allows developers to improve their productivity by reducing time and effort required for creating and maintaining appropriate DB sates, and enhances the quality of DB applications since they are conducive to a wider variety of test cases and support regression tests. Academically, this research deepens our understanding and introduces new approach to testing enterprise systems by analyzing patterns of SQL usages and defining a grammar to express and process the patterns.

• Maxillofacial Plastic and Reconstructive Surgery
• /
• v.28 no.2
• /
• pp.118-126
• /
• 2006

In oral and maxillofacial surgery, bone graft is very important procedure for functional and esthetic reconstruction. So, many researcher studied about bone graft material like autogenous bone, allograft bone and artificial bone materials. The purpose of this study is to evaluate the quantity of bone generation induced by $Grafton^{(R)}$ graft, human allogenic demineralized bone matrix. Total 24 sites of artificial bony defects prepared using trephin bur(diameter 8 mm) on parietal bone of six adult New Zealand White rabbits. Experimental group had six defect sites which grafted $Grafton^{(R)}$(0.1 cc). Active control group had nine defect sites, into which fresh autogenous bone harvested from own parietal bone was grafted and passive control group had nine defect sites without bone graft. After six weeks postoperatively, the rabbits were sacrificed. The defects and surrounding tissue were harvested and decalcified in 10% EDTA, 10% foamic-acid. Specimens were stained with H&E. New bone area percentage in whole defect area was measured by IMT(VT) image analysis program. Quantity of bone by $Grafton^{(R)}$ graft was smaller than that of autograft and larger than that of empty defects. In histologic view $Grafton^{(R)}$ graft site and autograft site showed similar healing progress but it was observed that newly formed bone in active control group was more mature. In empty defect, quantity and thickness of new bone formation was smaller than in $Grafton^{(R)}$-grafted defect. $Grafton^{(R)}$ is supposed to be a useful bone graft material instead of autogenous bone if proper maintenance for graft material stability and enough healing time were obtained.

• Asia Marketing Journal
• /
• v.12 no.2
• /
• pp.155-175
• /
• 2010

Most of the service industries including finance and telecommunications have become matured and saturated. The competitions have become severe while the differences among brands become smaller. Therefore maintaining good relationships with customers has been critical for the service providers. In case of credit card and debit card, the similar patterns are shown. It is important for them to maintain good relationships with customers, and therefore, they have used marketing program which provides customized services to customers and utilizes the membership programs. Not only do they build and maintain good relationships, but also highlight their brands from the emotional aspects. For example, KB Card or Hyundai Card uses well-known designers' works for their credit card design. As well, they differentiate the designs of credit cards to stress on their brand personalities. BC Card introduced the credit card with perfume that a customer would like. Even though the credit card is small and not shown to public easily, it becomes more important for those companies to touch the customers' feelings with the brand personalities and their images. This is partly because of changes in consumers' lifestyles. Y-generations becomes highly likely to express themselves in many different ways and more emotional than X-generations. For the Y-generations, therefore, even credit cards in the wallet should be personalized and well-designed. In line with it, credit cards with good design can be seen as an example of brand identity, where different design for each customer can be used to recognize the membership groups that customers want to belong. On the other hand, these credit card companies offer the special treatment benefits for those customers who are heavy users for the cards. For example, those customers who love sports will receive some special discounts when they use their credit cards for sports related products. Therefore this study attempted to explore the relationships between relational benefits, brand identification and loyalty. It has been well known that relational benefits and brand identification lead to loyalty independently from many other studies, but there has been few study to review all the three variables all together in a research model. Furthermore, as reviewed above, in the card industry, many companies attempt to associate the brand image with their products to fit their customers' lifestyles while relational benefits are still playing an important role for their business. Therefore in our research model, relational benefits, brand identification, and loyalty are all included. We focus on the mediating effect of brand identification. From the relational benefits perspective, only special treatment benefit and confidence benefit are included. Social benefit is not applicable for this credit card industry because not many cases of face-to-face interaction can be found. From the brand identification perspective, personal brand identity and social brand identity are reviewed and included in the model. Overall, the research model emphasizes that the relationships between relational benefits and loyalty will be mediated by the effect of brand identification. The effects of relational benefits which are confidence benefit and special treatment benefits on loyalty will be realized when they fit to the personal brand identity and social brand identity. In the research model, therefore, the relationships between confidence benefit and social brand identity, and between confidence benefit and personal identity are hypothesized while the effects of special treatment benefit on social brand identity and personal brand identity are hypothesized. Loyalty, then, is hypothesized to have positive relationships with personal brand identity and social brand identity. In addition, confidence benefit among the relational benefits is expected to have a direct, positive relationship with loyalty because confidence benefit has been recognized as a critical factor for good relationships and satisfaction. Data were collected from college students who have been using either credit cards or debit cards. College students were regarded good subjects because they are in Y-generation cohorts and have tendency to express themselves more. Total sample size was two hundred three at the beginning, but after deleting those data with many missing values, one hundred ninety-seven data points were remained and used for the model testing. Measurement items were brought from the previous literatures and modified for this research. To test the reliability, using SPSS 14, chronbach's α was examined and all the values were from .874 to .928 exceeding over .7. Using AMOS 7.0, confirmatory factor analysis was conducted to investigate the measurement model. The measurement model was found good fit with χ²(67)=188.388 (p= .000), GFI=.886, AGFI=.821, CFI=.941, RMSEA=.096. Using AMOS 7.0, structural equation modeling has been used to analyze the research model. Overall, the research model fit were χ²(68)=188.670 (p= .000), GFI=.886, AGFI=,824 CFI=.942, RMSEA=.095 indicating good fit. In details, all the paths hypothesized in the research model were found significant except for the path from social brand identity to loyalty. Personal brand identity leads to loyalty while both confidence benefit and special treatment benefit have a positive relationships with personal and social identities. As well, confidence benefit has a direct positive effect on loyalty. The results indicates the followings. First, personal brand identity plays an important role for credit/debit card usage. Therefore even for the products which are not shown to public easy, design and emotional aspect can be important to fit the customers' lifestyles. Second, confidence benefit and special treatment benefit have a positive effects on personal brand identity. Therefore it will be needed for marketers to associate the special treatment and trust and confidence benefits with personal image, personality and personal identity. Third, this study found again the importance of confidence and trust. However interestingly enough, social brand identity was not found to be significantly related to loyalty. It can be explained that the main sample of this study consists of college students. Those strategies to facilitate social brand identity are focused on high social status groups while college students have not been established their status yet.