• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2013년도 추계학술발표대회
• /
• pp.318-319
• /
• 2013

수년간 스마트폰, 패드 등 모바일 단말의 증가와 P2P, mVoIP, IPTV, Kakao Talk 등 신규 서비스들의 등장 및 이용으로 트래픽이 급증하고 있다. DPI방식의 장비들은 패킷을 분석하고 이를 통해 특정 사용자나 서비스에 대한 제어가 가능하다. 최근 많은 통신사들은 서비스 및 트래픽 관리를 위해 DPI 시스템을 도입하고 있다. 본 논문에서는 DPI 기술과 장비들에 관하여 살펴보고 관련 장비의 성능측정을 위한 메트릭에 대하여 살펴보기로 한다.

• 인지과학
• /
• 제34권3호
• /
• pp.197-226
• /
• 2023

텍스트 생성요약은 자연어처리의 과업 중 하나로 긴 텍스트의 내용을 보존하면서 짧게 축약된 요약문을 생성한다. 생성요약 과업의 특성 상 본문의 핵심내용을 요약문에서 보존하는 것은 매우 중요하다. 기존의 생성요약 방법론은 정답요약과의 어휘 중첩도(Lexical-Overlap)를 기반으로 본문의 내용과 유창성을 측정했다. ROUGE는 생성요약 요약모델의 평가지표로 많이 사용하는 어휘 중첩도 기반의 평가지표이다. 생성요약 벤치마크에서 ROUGE가 49점대로 매우 높은 성능을 보임에도 불구하고, 생성한 요약문과 본문의 내용이 불일치하는 경우가 30% 가량 존재한다. 본 연구에서는 정답요약의 도움 없이 본문만을 활용해 생성요약 모델의 성능을 평가하는 방법론을 제안한다. 본 연구에서 제안한 평가점수를 AggreFACT의 라벨과 상관도 분석결과, 다음의 두 가지 경우 가장 높은 상관관계를 보였다. 첫 번째는 Transformer 구조의 인코더-디코더 구조에 대규모 사전학습을 진행한 BART와 PEGASUS 등을 생성요약 모델의 베이스라인으로 사용한 경우이고, 두 번째는 요약문 전체에 걸쳐 오류가 발생한 경우이다.

• The Journal of Asian Finance, Economics and Business
• /
• 제6권3호
• /
• pp.235-245
• /
• 2019

This research focuses on the relationships among customer satisfaction, delight, and loyalty. Although customer delight is one facet of an affective evaluation that can be predicted from customer satisfaction as cognitive component of the evaluation, there is no empirical examination on the casual relationship among customer satisfaction, delight, and loyalty. This study aims at addressing this gap in the service literature. The research questions are (1) How is customer satisfaction related to customer delight? and (2) Does customer delight matter in the relationship between customer satisfaction and loyalty? Data from a survey of consumers across upscale restaurant and retail bank in Korea were obtained. Our results show that customer satisfaction contributes positively to customer delight, and that customer delight plays a significant role in the relationship between customer satisfaction and loyalty. This chained relationship from customer satisfaction to customer delight to customer loyalty suggests that achieving customer delight represents one of the underlying pathways through which basic or core requirements expected by customers are satisfied. Our finding suggests that service firms need to monitor and manage their levels of customer delight as a performance metric, and delighting customers may be an important strategy to build competitive advantage through customer loyalty.

• 한국멀티미디어학회논문지
• /
• 제15권10호
• /
• pp.1205-1211
• /
• 2012

대사성 골 질환인 골다공증(Osteoporosis)의 조기 진단을 위해 X 선 영상에서 골 밀도를 측정하는 방법이 최근 연구되고 있다. 골 밀도는 X 선 영상에서 뼈가 분리되고, 분리된 영역에서의 픽셀에 의해 BMD가 측정되는데, 개선된 영상에서의 정밀한 뼈 추출이 주요한 요소이므로 X 선 영상의 개선은 골다공증의 조기 진단을 위해 필수적이다. 본 논문에서는 sparse 표현을 도입하여 다중(multiple) 잡음을 갖는 X 선 영상을 개선시키는 방법을 제안한다. 실험을 통해 제안한 방법의 결과가 기존의 방법인 웨이블릿 BayesShrink 잡음 제거 방법 및 일반적 sparse 표현 모델의 잡음 제거 방법의 결과에 비해 개선됨을 CNR(Contrast to Noise Ratio) 및 cut-view를 통해 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권12호
• /
• pp.4269-4292
• /
• 2014

Currently, increasing numbers of access points (AP) are being deployed in enterprise offices, campuses and municipal downtowns for flexible Internet connectivity, but most of these access points are idle or redundant most of the time, which causes significant energy waste. Therefore, with respect to power conservation, applying energy efficient strategies in WIFI networks is strongly advocated. One feasible method is dynamically managing network resources, particularly APs, by powering devices on or off. However, when an AP is powered on, the device is initialized through a long boot time, during which period clients cannot be associated with it; therefore, the network performance would be greatly impacted. In this paper, based on a global view of an entire WLAN, we propose an AP selection technology, known as Temporary Access Selection (TAS). The criterion of TAS is a fusion metric consisting of two evaluation indexes which are based on throughput and battery life, respectively. TAS is both service and clients' preference specific through balancing the data rate, battery life and packet size. TAS also works well independently in traditional WLANs in which no energy efficient strategy is deployed. Moreover, this paper demonstrates the feasibility and performance of TAS through experiments and simulations with Network Simulator version 3 (NS3).

• 디지털산업정보학회논문지
• /
• 제5권4호
• /
• pp.87-97
• /
• 2009

Internet network routing system is used to prevent spread and distribution of malicious data traffic. This study is based on analysis of diagnostic weakness structure in the network security domain. We propose an improved integrated multi-level protection domain for in the internal route of groupware. This paper's protection domain is designed to handle the malicious data traffic in the groupware and finally leads to lighten the load of data traffic and improve network security in the groupware. Infrastructure of protection domain is transformed into five-stage blocking domain from two or three-stage blocking. Filtering and protections are executed for the entire server at the gateway level and internet traffic route ensures differentiated protection by dividing into five-stage. Five-stage multi-level network security domain's malicious data traffic protection performance is better than former one. In this paper, we use a trust evaluation metric for measuring the security domain's performance and suggested algorithm.

• 한국IT서비스학회지
• /
• 제11권1호
• /
• pp.211-221
• /
• 2012

Many metric-based classification models have been proposed to predict fault-proneness of software module. This paper presents two prediction models using Bayesian classifier which is one of the most popular modern classification algorithms. Bayesian model based on Bayesian probability theory can be a promising technique for software quality prediction. This is due to the ability to represent uncertainty using probabilities and the ability to partly incorporate expert's knowledge into training data. The two models, Na$\ddot{i}$veBayes(NB) and Bayesian Belief Network(BBN), are constructed and dimensionality reduction of training data and test data are performed before model evaluation. Prediction accuracy of the model is evaluated using two prediction error measures, Type I error and Type II error, and compared with well-known prediction models, backpropagation neural network model and support vector machine model. The results show that the prediction performance of BBN model is slightly better than that of NB. For the data set with ambiguity, although the BBN model's prediction accuracy is not as good as the compared models, it achieves better performance than the compared models for the data set without ambiguity.

• 한국해양공학회지
• /
• 제34권6호
• /
• pp.442-453
• /
• 2020

Underwater color images suffer from low visibility and color cast effects caused by light attenuation by water and floating particles. This study applied single image enhancement techniques to enhance the quality of underwater images and compared their performance with real underwater images taken in Korean waters. Dark channel prior (DCP), gradient transform, image fusion, and generative adversarial networks (GAN), such as cycleGAN and underwater GAN (UGAN), were considered for single image enhancement. Their performance was evaluated in terms of underwater image quality measure, underwater color image quality evaluation, gray-world assumption, and blur metric. The DCP saturated the underwater images to a specific greenish or bluish color tone and reduced the brightness of the background signal. The gradient transform method with two transmission maps were sensitive to the light source and highlighted the region exposed to light. Although image fusion enabled reasonable color correction, the object details were lost due to the last fusion step. CycleGAN corrected overall color tone relatively well but generated artifacts in the background. UGAN showed good visual quality and obtained the highest scores against all figures of merit (FOMs) by compensating for the colors and visibility compared to the other single enhancement methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권10호
• /
• pp.4968-4986
• /
• 2017

In this paper, we primarily address the difficulty of automatic generation of a plausible depth map from a single image in an unstructured environment. The aim is to extrapolate a depth map with a more correct, rich, and distinct depth order, which is both quantitatively accurate as well as visually pleasing. Our technique, which is fundamentally based on a preexisting DepthTransfer algorithm, transfers depth information at the level of superpixels. This occurs within a framework that replaces a pixel basis with one of instance-based learning. A vital superpixels feature enhancing matching precision is posterior incorporation of predictive semantic labels into the depth extraction procedure. Finally, a modified Cross Bilateral Filter is leveraged to augment the final depth field. For training and evaluation, experiments were conducted using the Make3D Range Image Dataset and vividly demonstrate that this depth estimation method outperforms state-of-the-art methods for the correlation coefficient metric, mean log10 error and root mean squared error, and achieves comparable performance for the average relative error metric in both efficacy and computational efficiency. This approach can be utilized to automatically convert 2D images into stereo for 3D visualization, producing anaglyph images that are visually superior in realism and simultaneously more immersive.

• 정보보호학회논문지
• /
• 제34권1호
• /
• pp.53-60
• /
• 2024

안전한 네트워크를 위해 보안 평가는 필수불가결한 과정으로, 위험을 관리하기 위해서는 적절한 성능지표가 있어야 한다. 가장 널리 사용하고 있는 정량적 지표로는 CVSS가 있다. CVSS는 주관성과 해석의 복잡성, 보안 위험의 관점에서 맥락을 고려하지 못한다는 문제가 있다. 이러한 문제를 보완하기 위해 ISO/IEC 15408 문서의 보안 개념 및 관계도를 바탕으로 공격자, 위협, 대응, 자산의 4가지를 항목화하고 수치화 하는 지표를 제안한다. 네트워크 스캐닝을 통해 발견된 취약점은 약점, 공격패턴의 연결관계에 의해 MITRE ATT&CK의 기술과 매핑시킬 수 있다. 우리는 MITRE ATT&CK 의 Groups, Tactic, Mitigations을 이용하여 일관성을 가지며 직관적인 점수를 산출한다. 이에 따라 보안 평가 관리자가 다양한 관점의 보안지표 중 선택할 수 있는 폭을 넓히고, 사이버 네트워크의 보안을 강화하는데 긍정적인 영향을 기대한다.