• 방송공학회논문지
• /
• 제17권2호
• /
• pp.223-236
• /
• 2012

본 논문에서는 인간 인지 기반 비디오 코딩을 위한 비디오 처리 방법을 개발한다. 제안하는 방법은 율-왜곡(rate-distortion) 최적화의 영향뿐만 아니라 제한적인 시, 공간 해상도, 지역적인 움직임 이력(history), visual saliency에 의한 인간 시각 인지를 고려한다. 이러한 인간의 인지적인 효과들을 고려하기 위하여 본 논문에서는 움직임 패턴을 모델링하고 Hedge 알고리듬을 사용하여 움직임 패턴을 결정하는 기법을 개발한다. 그 다음, 제안한 움직임 패턴과 기존의 visual saliency와의 결합을 통하여 인간 시각 인지 모델을 수립한다. 제안된 인간 시각 인지 모델을 구현하기 위하여 기존의 foveation filtering 방법을 확장한다. 시각적 자극이 덜한 지역만을 부드럽게(smoothing)하는 기존의 foveation filtering 기법과 비교하여 제안하는 foveation filtering 기법은 인간 시각 인지 모델에 따라 지역적으로 부드럽게 또는 지역적 특성을 향상시킴으로써, 시각적 자극이 덜한 지역에서 줄여진 대역폭을 효과적으로 시각적 자극이 큰 지역에서 사용하도록 이동 시킬 수 있는 장점이 있다. 제안된 방법의 성능은 전반적인 비디오 화질을 만족할 뿐만 아니라 인간이 인지하는 화질의 품질을 12%~44% 향상시킨다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권2호
• /
• pp.483-503
• /
• 2014

In this paper, human action recognition using pyramid histograms of oriented gradients and collaborative multi-task learning is proposed. First, we accumulate global activities and construct motion history image (MHI) for both RGB and depth channels respectively to encode the dynamics of one action in different modalities, and then different action descriptors are extracted from depth and RGB MHI to represent global textual and structural characteristics of these actions. Specially, average value in hierarchical block, GIST and pyramid histograms of oriented gradients descriptors are employed to represent human motion. To demonstrate the superiority of the proposed method, we evaluate them by KNN, SVM with linear and RBF kernels, SRC and CRC models on DHA dataset, the well-known dataset for human action recognition. Large scale experimental results show our descriptors are robust, stable and efficient, and outperform the state-of-the-art methods. In addition, we investigate the performance of our descriptors further by combining these descriptors on DHA dataset, and observe that the performances of combined descriptors are much better than just using only sole descriptor. With multimodal features, we also propose a collaborative multi-task learning method for model learning and inference based on transfer learning theory. The main contributions lie in four aspects: 1) the proposed encoding the scheme can filter the stationary part of human body and reduce noise interference; 2) different kind of features and models are assessed, and the neighbor gradients information and pyramid layers are very helpful for representing these actions; 3) The proposed model can fuse the features from different modalities regardless of the sensor types, the ranges of the value, and the dimensions of different features; 4) The latent common knowledge among different modalities can be discovered by transfer learning to boost the performance.

• 대한인간공학회지
• /
• 제29권6호
• /
• pp.907-914
• /
• 2010

Cognitive model, that is cognitive architecture, is the model expressed with computer program to show the process how human solve a certain problem and it is continuously under investigation through various fields of study such as cognitive engineering, computer engineering, and cognitive psychology. In addition, the much extensive applicability of cognitive model usually helps it to be used for quantitative prediction of human Behavior or Natural programming of human performance in many HCI areas including User Interface Usability, artificial intelligence, natural programming language and also Robot engineering. Meanwhile, when a system designed, an usability test about conceptual design of interface is needed and in this case, analysis evaluation using cognitive model like GOMS or ACT-R is much more effective than empirical evaluation which naturally needs products and subjects. In particular, if we consider the recent trend of very short-end term between a previous technology development and the next new one, it would take time and much efforts to choose subjects and train them in order to conduct usability test which is repeatedly followed in the process of a system development and this finally would bring delays of development of a new system. In this study, we predicted quantitatively the human behavior processes which contains cognitive processes for menu selection in touch screen interface through ACT-R, one of the common method of usability test. Throughout the study, it was shown that the result using cognitive model was equal with the result using existing empirical evaluation. And it is expected that cognitive model has a possibility not only to be used as an effective methodology for evaluation of HCI products or system but also to contribute the activation of HCI cognitive modeling in Korea.

• 융합보안논문지
• /
• 제23권5호
• /
• pp.65-72
• /
• 2023

본 논문에서는 LSTM(Long Short-Term Memory)을 기반으로 하는 Deep Learning 모델을 구축하여 인간의 습관적 특성을 고려한 악성 도메인 탐지 방법을 제시한다. DGA(Domain Generation Algorithm) 악성 도메인은 인간의 습관적인 실수를 악용하여 심각한 보안 위협을 초래한다. 타이포스쿼팅을 통한 악성 도메인의 변화와 은폐 기술에 신속히 대응하고, 정확하게 탐지하여 보안 위협을 최소화하는 것이 목표이다. LSTM 기반 Deep Learning 모델은 악성코드별 특징을 분석하고 학습하여, 생성된 도메인을 악성 또는 양성으로 자동 분류한다. ROC 곡선과 AUC 정확도를 기준으로 모델의 성능 평가 결과, 99.21% 이상 뛰어난 탐지 정확도를 나타냈다. 이 모델을 활용하여 악성 도메인을 실시간 탐지할 수 있을 뿐만 아니라 다양한 사이버 보안 분야에 응용할 수 있다. 본 논문은 사용자 보호와 사이버 공격으로부터 안전한 사이버 환경 조성을 위한 새로운 접근 방식을 제안하고 탐구한다.

• 산업경영시스템학회지
• /
• 제41권2호
• /
• pp.65-73
• /
• 2018

The dynamic capabilities of sensing market signals, creating new opportunities and reconfiguring resources and capabilities to new opportunities in a rapidly changing economic environment determines the competitiveness of the enterprise to create added value and survival. This study conceptualized a two-stage performance measurement framework based on the casual model of resource (input)-process-performance (output). We have developed a 'Process capability index' that reflect the dynamic capabilities factors as a key intermediary product linking resource inputs and performance outputs in enterprise performance measurement. The process capability index consists of four elements : manpower (level of human resource), operation productivity, structure and risk management. The DEA (Data Envelopment Analysis) model was applied to the developed performance indicators to analyze the branch office performance of a telecom company. Process capability efficiency (stage 1) uses resource inputs to reach a certain level of process capabilities. In performance result efficiency (stage 2), the process capabilities are used to generate sales revenues and subscribers. The two-stage DEA model derives intermediate output values that optimize the individual stages simultaneously. Some branch offices in the telecom company have focused on process capability efficiency or some other branch offices focused on performance result efficiency. Positioning map using two-stage efficiency decomposition and benchmarking can help identify the sources of inefficiencies and visualize strategic directions for performance optimization. Applications of two-stage DEA in conjunction with the case study that are meaningfully used in performance measurement areas have been scarce. In particular, this paper has the contribution to present a new performance measurement model considering the organization theory, the dynamic capabilities.

• 한국위성정보통신학회논문지
• /
• 제12권1호
• /
• pp.28-33
• /
• 2017

본 논문에서는 여러 곳에서 통용되고 있는 기존의 CW(continuous-wave) 레이더 송수신모델을 수정한 송수신모델을 제안한다. 최근에 심장박동과 호흡을 검출하기 위해 기존신호모델을 기반으로 CW(continuous-wave) 레이더에 대한연구가 진행 되고 있다. 그러나 이 통용되고 있는 수신모델은 인체 공학적으로 개념과 일치하지 않기 때문에 이모델을 근거로 실험을 할 경우 실제 개발되는 시스템 성능을 정확하게 예측할수 없다. 본 논문에서는 인체 공학상 개념과 일치하는 수정된 CW(continuous-wave) 레이더 송수신모델을 제안하고 가운시안 잡음rhk 다중경로 환경에서 심장박동 및 호흡검출에 대한 시뮬레이션을 수행하고 이를 기존수신모델과의 심장박동과 호흡검출에 대한 성능을 비교 분석한다.

• ETRI Journal
• /
• 제46권1호
• /
• pp.22-34
• /
• 2024

Exposure to varied noisy environments impairs the recognition performance of artificial intelligence-based speech recognition technologies. Degraded-performance services can be utilized as limited systems that assure good performance in certain environments, but impair the general quality of speech recognition services. This study introduces an audiovisual speech recognition (AVSR) model robust to various noise settings, mimicking human dialogue recognition elements. The model converts word embeddings and log-Mel spectrograms into feature vectors for audio recognition. A dense spatial-temporal convolutional neural network model extracts features from log-Mel spectrograms, transformed for visual-based recognition. This approach exhibits improved aural and visual recognition capabilities. We assess the signal-to-noise ratio in nine synthesized noise environments, with the proposed model exhibiting lower average error rates. The error rate for the AVSR model using a three-feature multi-fusion method is 1.711%, compared to the general 3.939% rate. This model is applicable in noise-affected environments owing to its enhanced stability and recognition rate.

• 한국정보시스템학회지:정보시스템연구
• /
• 제14권1호
• /
• pp.125-147
• /
• 2005

This research focused on the comparison of corporate business mokels to explain different financial performances on the eBusiness domain. Especially, because customers have the more buying [ower tha other business areas, customer value and the service quality were prepared for independent variables and operational margin which can be obtained by publicize report was used for a dependent variable in stead of the other variables dependent on human perception as well. As a result, this research found that the customer value measured by service quality concept impact on the financial performance of eBusiness corporation positively. To find out more delicate results, structural equation was used for statistical method using 324 survey samples on 10 corporations. Though data using for statistical analysis were divided into individual and corporate level and have the time gap between research time and financial performance publicized period, the value of this research is that the customer value and service quality concepts with very objective financial information were input for constructing a research model.

• 대한인간공학회지
• /
• 제19권3호
• /
• pp.39-49
• /
• 2000

During human-computer interaction(HCI), people typically send inputs to computers through electromechanical pointing devices. Many applied studies have therefore evaluated cursor-positioning movements made with various pointing devices. Though there were so many studies about performance of various pointing devices, it was nearly impossible to compare device performance each other until the Fitts' law was applied. It does appear that Fitts' law may predict performance reasonably well for the one C-D gain level. But in varying C-D gain levels, Fitts' law could not predict movement time. This study investigated the effects of C-D gain in mouse movement time and suggested a revised Fitts' model including C-D gain as an independent variable. The revised Fitts' model may use to measure the performance of various devices in varying C-D gain levels.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2008년도 학술대회 1부
• /
• pp.487-492
• /
• 2008

Facial expression recognition is an intensive research area for designing Human Computer Interfaces. In this work, we present a new facial expression recognition system utilizing Enhanced Independent Component Analysis (EICA) for feature extraction and discrete Hidden Markov Model (HMM) for recognition. Our proposed approach for the first time deals with sequential images of emotion-specific facial data analyzed with EICA and recognized with HMM. Performance of our proposed system has been compared to the conventional approaches where Principal and Independent Component Analysis are utilized for feature extraction. Our preliminary results show that our proposed algorithm produces improved recognition rates in comparison to previous works.