• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권9호
• /
• pp.3258-3273
• /
• 2021

Malware is a severe threat to the computing system and there's a long history of the battle between malware detection and anti-detection. Most traditional detection methods are based on static analysis with signature matching and dynamic analysis methods that are focused on sensitive behaviors. However, the usual detections have only limited effect when meeting the development of malware, so that the manual update for feature sets is essential. Besides, most of these methods match target samples with the usual feature database, which ignored the characteristics of the sample itself. In this paper, we propose a new malware detection method that could combine the features of a single sample and the general features of malware. Firstly, a structure of Directed Cyclic Graph (DCG) is adopted to extract features from samples. Then the sensitivity of each API call is computed with Markov Chain. Afterward, the graph is merged with the chain to get the final features. Finally, the detectors based on machine learning or deep learning are devised for identification. To evaluate the effect and robustness of our approach, several experiments were adopted. The results showed that the proposed method had a good performance in most tests, and the approach also had stability with the development and growth of malware.

• 한국포장학회지
• /
• 제28권3호
• /
• pp.237-244
• /
• 2022

Onions are a significant crop in Korea, and cultivation is increasing every year along with high demand. Onions are planted in the fall and mainly harvested in June, the rainy season, therefore, physiological changes in onion bulbs during long-term storage might have happened. Onions are stored in cold room and at adequate relative humidity to avoid quality loss. In this study, bio-yield stress and weight loss were measured as the quality parameters of net packaged onions during 10 weeks of storage, and the storage environmental conditions are monitored using sensor networks systems. Quality estimation of net packaged onion during storage was performed using the storage environmental condition data through machine learning approaches. Among the suggested estimation models, support vector regression method showed the best accuracy for the quality estimation of net packaged onions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권8호
• /
• pp.2473-2489
• /
• 2022

Urban development has brought about the increasing saturation of urban traffic demand, and traffic congestion has become the primary problem in transportation. Roads are in a state of waiting in line or even congestion, which seriously affects people's enthusiasm and efficiency of travel. This paper mainly studies the discrete domain path planning method based on the flow data. Taking the traffic flow data based on the highway network structure as the research object, this paper uses the deep learning theory technology to complete the path weight determination process, optimizes the path planning algorithm, realizes the vehicle path planning application for the expressway, and carries on the deployment operation in the highway company. The path topology is constructed to transform the actual road information into abstract space that the machine can understand. An appropriate data structure is used for storage, and a path topology based on the modeling background of expressway is constructed to realize the mutual mapping between the two. Experiments show that the proposed method can further reduce the interpolation error, and the interpolation error in the case of random missing is smaller than that in the other two missing modes. In order to improve the real-time performance of vehicle path planning, the association features are selected, the path weights are calculated comprehensively, and the traditional path planning algorithm structure is optimized. It is of great significance for the sustainable development of cities.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 춘계학술대회
• /
• pp.204-206
• /
• 2021

실시간 시추매개변수 예측은 시추효율의 극대화 관점에서 상당히 중요한 연구이다. 시추 극대화 방법 중 시추속도를 향상시키는 방법이 일반적인데 이는 굴진율, 시추스트링 회전속도, 비트 하중, 시추이수 유량과 연관관계를 지니고 있다. 본 연구는 실시간 시추매개변수 중 하나인 굴진율을 순환신경망기반 딥러닝 모델을 이용하여 예측하는 방법을 제안하였으며 기존의 물리적 기반의 굴진율 모델과 딥러닝 모델을 이용한 예측 모델을 비교해 보고자 한다.

• 한국융합학회논문지
• /
• 제9권10호
• /
• pp.21-28
• /
• 2018

학습코스 선정에 많은 어려움과 시행착오를 겪고 있는 사용자들에게 수준별 학습코스를 제공하기 위해, ITS(Intelligence Tutoring System)를 위한 동적인 학습자 맞춤형 학습코스 추천 모델을 개발하였다. 이를 위해, 개인화 학습코스 추천모델에서는 먼저 학습자 프로파일을 분석하고, 단어별 가중치를 계산하여 핵심 키워드를 추출한다. 추출된 단어는 Cosine Similarity 기법을 통해 유사도를 측정하고, 최종적으로 유사도가 높은 상위 3개 과정이 학습자에게 추천된다. 추천모델의 효과를 분석하기 위해, 경기도 소재 교육기관에 추천모델을 적용하였고, 만족도 조사를 통하여 설문 항목별 평균, 표준편차, 왜도, 첨도 값을 계산하였다. 실험결과, 정확성, 새로움, 자기참조, 유용성에서 높은 만족도를 보였으며, 추천모델의 실효성을 검증했다. 본 연구는 그동안 국내 외에서 충분히 다뤄지지 않았던 기계학습 중심의 맞춤형 학습코스를 추천했다는 점에서 의미가 있다.

• 한국콘텐츠학회논문지
• /
• 제18권11호
• /
• pp.447-454
• /
• 2018

본 논문은 컨볼루션 신경망으로 이루어진 학습 모델을 통해 스테레오 영상의 깊이영상 생성 알고리즘을 제안한다. 제안하는 알고리즘은 좌, 우 시차 영상을 입력으로 받아 각 시차영상의 주요 특징을 추출하는 특징 추출부와 추출된 특징을 이용하여 시차 정보를 학습하는 깊이 학습부로 구성된다. 우선 특징 추출부는 2D CNN 계층들로 이루어진 익셉션 모듈(xception module) 및 ASPP 모듈(atrous spatial pyramid pooling) module을 통해 각각의 시차영상에 대한 특징맵을 추출한다. 그 후 각 시차에 대한 특징 맵을 시차에 따라 3차원 형태로 쌓아 3D CNN을 통해 깊이 추정 가중치를 학습하는 깊이 학습부를 거친 후 깊이 영상을 추정한다. 제안하는 알고리즘은 객체 영역에 대해 기존의 다른 학습 알고리즘들 보다 정확한 깊이를 추정하였다.

• 한국산학기술학회논문지
• /
• 제22권2호
• /
• pp.714-720
• /
• 2021

머신 러닝 및 신호처리에 활용되고 있는 정보이론적 학습법(ITL, information theoretic learning)은 커널 사이즈(σ) 설정이 매우 민감한 어려움을 지닌다. ITL의 성능지표중 하나인 코렌트로피 함수를 최대화하는 성능지표에 대해, 기울기에 존재하는 1/σ2를 제거한 뒤 남은 커널 사이즈에 대해 적응적으로 조절하는 방법들이 연구되었다. 이 논문에서는, 1/σ2의 커널 사이즈가 실제 시스템의 민감성이나 불안정에 큰 역할을 하고 있으며 남은 부분에 존재하는 커널 사이즈에 대한 최적해는 오차의 절대값 근방에 수렴함에 따라 오히려 수렴 후 가중치 갱신을 멈추게 하는 부작용이 나타남을 밝혔다. 이에 적응적 커널 사이즈 조절 대신 적절한 상수를 선택하는 것이 보다 효과적이라는 것을 제안하였고, 실험결과에서 동일한 수렴 속도에 약 2dB 향상된 정상상태 MSE를 보였다. 제안한 방식을 더욱 열악한 다경로 채널환경에 적용하여 실험한 결과 4dB 이상의 성능향상을 보여 제안한 방식은 열악한 상황일수록 더욱 향상된 성능을 보임을 알 수 있다.

• 대한건축학회논문집:구조계
• /
• 제36권3호
• /
• pp.113-120
• /
• 2020

The purpose of this study is to compare the models of Deep Learning-based Convolution Neural Network(CNN) for concrete crack detection. The comparison models are AlexNet, GoogLeNet, VGG16, VGG19, ResNet-18, ResNet-50, ResNet-101, and SqueezeNet which won ImageNet Large Scale Visual Recognition Challenge(ILSVRC). To train, validate and test these models, we constructed 3000 training data and 12000 validation data with 256×256 pixel resolution consisting of cracked and non-cracked images, and constructed 5 test data with 4160×3120 pixel resolution consisting of concrete images with crack. In order to increase the efficiency of the training, transfer learning was performed by taking the weight from the pre-trained network supported by MATLAB. From the trained network, the validation data is classified into crack image and non-crack image, yielding True Positive (TP), True Negative (TN), False Positive (FP), False Negative (FN), and 6 performance indicators, False Negative Rate (FNR), False Positive Rate (FPR), Error Rate, Recall, Precision, Accuracy were calculated. The test image was scanned twice with a sliding window of 256×256 pixel resolution to classify the cracks, resulting in a crack map. From the comparison of the performance indicators and the crack map, it was concluded that VGG16 and VGG19 were the most suitable for detecting concrete cracks.

• 한국멀티미디어학회논문지
• /
• 제25권10호
• /
• pp.1375-1385
• /
• 2022

In this paper, we propose a semi-supervised domain adaptation solution to deal with practical face recognition (FR) scenarios where a single face image for each target identity (to be recognized) is only available in the training phase. Main goal of the proposed method is to reduce the discrepancy between the target and the source domain face images, which ultimately improves FR performances. The proposed method is based on the Domain Adatation network (DAN) using an MMD loss function to reduce the discrepancy between domains. In order to train more effectively, we develop a novel loss function learning strategy in which MMD loss and cross-entropy loss functions are adopted by using different weights according to the progress of each epoch during the learning. The proposed weight adoptation focuses on the training of the source domain in the initial learning phase to learn facial feature information such as eyes, nose, and mouth. After the initial learning is completed, the resulting feature information is used to training a deep network using the target domain images. To evaluate the effectiveness of the proposed method, FR performances were evaluated with pretrained model trained only with CASIA-webface (source images) and fine-tuned model trained only with FERET's gallery (target images) under the same FR scenarios. The experimental results showed that the proposed semi-supervised domain adaptation can be improved by 24.78% compared to the pre-trained model and 28.42% compared to the fine-tuned model. In addition, the proposed method outperformed other state-of-the-arts domain adaptation approaches by 9.41%.

• 공학교육연구
• /
• 제25권3호
• /
• pp.3-10
• /
• 2022

An educator may utilize various pedagogies such as hands-on practice, practicum, and project-based learning to enhance a student's competency. Among various pedagogies, project-based learning is one of the well-known pedagogies that may provide similar on-the-job experience. In general, an educator may divide the students into small groups and assign tasks to check students' cooperation skills and achievements during project-based learning. However, an educator may experience difficulties operating project-based learning due to the COVID-19 pandemic. Most of the team activities are done in non-face-to-face meetings, and as a consequence, the educator may not find an underachieving team easily and cannot intervene appropriately. This study introduces a rigorous analysis method to evaluate team activities to analyze individual students' participation and contributions. First, this study develops evaluation rubrics by conducting questionnaires to professors and students to find an appropriate weight value for the evaluation scale. Then, this research introduces an analysis environment to evaluate students automatically. The analysis environment collects dialog data from social network services and measures interactions among students in a team. After measuring interactions, the environment generates a report to visualize the team activities. We applied the proposed method and environment to the capstone design course to show the effectiveness of the method and environment. Based on the case study, the environment showed that the analysis method could easily indicate the teams' activities and check the level of participation.