• Journal of the Korea Society for Simulation
• /
• v.31 no.2
• /
• pp.11-20
• /
• 2022

There is a critical need for AI assistance in guard operations of Army base perimeters, which is exacerbated by changes in the national defense and security environment such as force reduction. In addition, the possibility for human error inherent to perimeter guard operations attests to the need for an innovative revamp of current systems. The purpose of this study is to propose a real-time object detection AI tailored to military CCTV surveillance with three unique characteristics. First, training data suitable for situations in which relatively small objects must be recognized is used due to the characteristics of military CCTV. Second, we utilize a data augmentation algorithm suited for military context applied in the data preparation step. Third, a noise reduction algorithm is applied to account for military-specific situations, such as camouflaged targets and unfavorable weather conditions. The proposed system has been field-tested in a real-world setting, and its performance has been verified.

• Journal of Digital Convergence
• /
• v.17 no.9
• /
• pp.71-78
• /
• 2019

This study aims to identify good borrowers within the context of P2P lending. P2P lending is a growing platform that allows individuals to lend and borrow money from each other. Inherent in any loans is credit risk of borrowers and needs to be considered before any lending. Specifically in the context of P2P lending, traditional models fall short and thus this study aimed to rectify this as well as explore the problem of class imbalances seen within credit risk data sets. This study implemented an over-sampling technique known as Synthetic Minority Over-sampling Technique (SMOTE). To test our approach, we implemented five benchmarking classifiers such as support vector machines, logistic regression, k-nearest neighbor, random forest, and deep neural network. The data sample used was retrieved from the publicly available LendingClub dataset. The proposed SMOTE revealed significantly improved results in comparison with the benchmarking classifiers. These results should help actors engaged within P2P lending to make better informed decisions when selecting potential borrowers eliminating the higher risks present in P2P lending.

• Journal of Broadcast Engineering
• /
• v.26 no.6
• /
• pp.748-757
• /
• 2021

In this paper, we propose a gaze estimation network in which eye landmark position detection and gaze direction vector estimation are integrated into one deep learning network. The proposed network uses the Stacked Hourglass Network as a backbone structure and is largely composed of three parts: a landmark detector, a feature map extractor, and a gaze direction estimator. The landmark detector estimates the coordinates of 50 eye landmarks, and the feature map extractor generates a feature map of the eye image for estimating the gaze direction. And the gaze direction estimator estimates the final gaze direction vector by combining each output result. The proposed network was trained using virtual synthetic eye images and landmark coordinate data generated through the UnityEyes dataset, and the MPIIGaze dataset consisting of real human eye images was used for performance evaluation. Through the experiment, the gaze estimation error showed a performance of 3.9, and the estimation speed of the network was 42 FPS (Frames per second).

• Journal of Internet Computing and Services
• /
• v.21 no.4
• /
• pp.17-23
• /
• 2020

Biometric information indicating measurement items related to human characteristics has attracted great attention as security technology with high reliability since there is no fear of theft or loss. Among these biometric information, fingerprints are mainly used in fields such as identity verification and identification. If there is a problem such as a wound, wrinkle, or moisture that is difficult to authenticate to the fingerprint image when identifying the identity, the fingerprint expert can identify the problem with the fingerprint directly through the preprocessing step, and apply the image processing algorithm appropriate to the problem. Solve the problem. In this case, by implementing artificial intelligence software that distinguishes fingerprint images with cuts and wrinkles on the fingerprint, it is easy to check whether there are cuts or wrinkles, and by selecting an appropriate algorithm, the fingerprint image can be easily improved. In this study, we developed a total of 17,080 fingerprint databases by acquiring all finger prints of 1,010 students from the Royal University of Cambodia, 600 Sokoto open data sets, and 98 Korean students. In order to determine if there are any injuries or wrinkles in the built database, criteria were established, and the data were validated by experts. The training and test datasets consisted of Cambodian data and Sokoto data, and the ratio was set to 8: 2. The data of 98 Korean students were set up as a validation data set. Using the constructed data set, five CNN-based architectures such as Classic CNN, AlexNet, VGG-16, Resnet50, and Yolo v3 were implemented. A study was conducted to find the model that performed best on the readings. Among the five architectures, ResNet50 showed the best performance with 81.51%.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.07a
• /
• pp.701-704
• /
• 2020

이미지 초해상도는 딥러닝의 발전과 함께 이를 활용하며 눈에 띄는 성능향상을 이루었다. 딥러닝을 기반으로 한 대부분의 이미지 초해상도 연구는 딥러닝 네트워크 모델의 구조에 대한 연구 위주로 진행되어 왔다. 그러나 최근 들어 딥러닝 기반의 이미지 초해상도가 합성된 데이터에 대해서는 높은 성능을 보이지만 실제 데이터에 대해서는 높은 성능을 보이지 못한다는 사실이 주목받고 있다. 이에 따라 모델 구조를 바꿔 성능을 향상 시키는 것에는 한계가 있어 데이터의 활용이나 학습 방법에 대한 연구의 필요성이 증대되고 있다. 따라서 본 논문은 이미지 초해상도를 위한 난이도 조절 기반 전이학습법(transfer learning)을 제안한다. 제안된 방법에서는 이미지 초해상도를 배율을 난이도가 쉬운 낮은 배율부터 순차적으로 전이학습을 진행한다. 이는 이미지 초해상도의 배율이 높아질수록 학습이 어렵기 때문이다. 결과적으로 본 논문에서는 높은 배율의 이미지 초해상도를 진행하기 위해 낮은 배율의 이미지 초해상도, 즉 난이도가 쉬운 학습부터 점진적으로 학습을 진행하였을 때 더욱 빠르고 효과적으로 학습할 수 있음을 보여준다. 제안된 전이학습 방법을 통해 적은 횟수의 업데이트로 학습을 진행하였을 때 일반적인 학습방법 대비 약 0.18 dB 의 PSNR 상승을 얻어, RealSR [9] 데이터셋에서 28.56 dB의 성능으로 파라미터 수 대비 높은 성능을 얻을 수 있었다.

• Journal of Intelligence and Information Systems
• /
• v.29 no.3
• /
• pp.419-437
• /
• 2023

As the number of thin filers in Korea surpasses 12 million, there is a growing interest in enhancing the accuracy of assessing their credit default risk to generate additional revenue. Specifically, researchers are actively pursuing the development of default prediction models using machine learning and deep learning algorithms, in contrast to traditional statistical default prediction methods, which struggle to capture nonlinearity. Among these efforts, Graph Neural Network (GNN) architecture is noteworthy for predicting default in situations with limited data on thin filers. This is due to their ability to incorporate network information between borrowers alongside conventional credit-related data. However, prior research employing graph neural networks has faced limitations in effectively handling diverse categorical variables present in credit information. In this study, we introduce the Transformer embedded Graph Convolutional Network (TeGCN), which aims to address these limitations and enable effective default prediction for thin filers. TeGCN combines the TabTransformer, capable of extracting contextual information from categorical variables, with the Graph Convolutional Network, which captures network information between borrowers. Our TeGCN model surpasses the baseline model's performance across both the general borrower dataset and the thin filer dataset. Specially, our model performs outstanding results in thin filer default prediction. This study achieves high default prediction accuracy by a model structure tailored to characteristics of credit information containing numerous categorical variables, especially in the context of thin filers with limited data. Our study can contribute to resolving the financial exclusion issues faced by thin filers and facilitate additional revenue within the financial industry.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.5
• /
• pp.162-177
• /
• 2020

The perception of traffic environment based on various sensors in autonomous driving system has a direct relationship with driving safety. Recently, as the perception model based on deep neural network is used due to the development of machine learning/in-depth neural network technology, a the perception model training and high quality of a training dataset are required. However, there are several realistic difficulties to collect data on all situations that may occur in self-driving. The performance of the perception model may be deteriorated due to the difference between the overseas and domestic traffic environments, and data on bad weather where the sensors can not operate normally can not guarantee the qualitative part. Therefore, it is necessary to build a virtual road environment in the simulator rather than the actual road to collect the traning data. In this paper, a training dataset collection process is suggested by diversifying the weather, illumination, sensor position, type and counts of vehicles in the simulator environment that simulates the domestic road situation according to the domestic situation. In order to achieve better performance, the authors changed the domain of image to be closer to due diligence and diversified. And the performance evaluation was conducted on the test data collected in the actual road environment, and the performance was similar to that of the model learned only by the actual environmental data.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.1081-1088
• /
• 2020

In this paper, the method of detecting sea fog through CCTV image is proposed based on convolutional neural networks. The study data randomly extracted 1,0004 images, sea-fog and not sea-fog, from a total of 11 ports or beaches (Busan Port, Busan New Port, Pyeongtaek Port, Incheon Port, Gunsan Port, Daesan Port, Mokpo Port, Yeosu Gwangyang Port, Ulsan Port, Pohang Port, and Haeundae Beach) based on 1km of visibility. 80% of the total 1,0004 datasets were extracted and used for learning the convolutional neural network model. The model has 16 convolutional layers and 3 fully connected layers, and a convolutional neural network that performs Softmax classification in the last fully connected layer is used. Model accuracy evaluation was performed using the remaining 20%, and the accuracy evaluation result showed a classification accuracy of about 96%.

• Smart Media Journal
• /
• v.10 no.4
• /
• pp.45-54
• /
• 2021

The skin is the body's first line of defense against external infection. When a skin disease strikes, the skin's protective role is compromised, necessitating quick diagnosis and treatment. Recently, as artificial intelligence has advanced, research for technical applications has been done in a variety of sectors, including dermatology, to reduce the rate of misdiagnosis and obtain quick treatment using artificial intelligence. Although previous studies have diagnosed skin diseases with low incidence, this paper proposes a method to classify common illnesses such as warts and corns using a convolutional neural network. The data set used consists of 3 classes and 2,515 images, but there is a problem of lack of training data and class imbalance. We analyzed the performance using a deep metric loss function and a cross-entropy loss function to train the model. When comparing that in terms of accuracy, recall, F1 score, and accuracy, the former performed better.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.1-9
• /
• 2023

Stress is a significant issue in modern society, often triggered by external or internal factors that are difficult to manage. When high stress persists over a long term, it can develop into a chronic condition, negatively impacting health and overall well-being. However, it is challenging for individuals experiencing chronic stress to recognize their condition, making early detection and management crucial. Using biosignals measured from wearable devices to detect stress could lead to more effective management. However, there are two main problems with using biosignals: first, manually extracting features from these signals can introduce bias, and second, the performance of classification models can vary greatly depending on the subject of the experiment. This paper proposes a model that reduces bias using convo utional autoencoders, which can represent the key features of data, and enhances generalizability by employing soft voting, a method of ensemble learning, to minimize performance variability. To verify the generalization performance of the model, we evaluate it using LOSO cross-validation method. The model proposed in this paper has demonstrated superior accuracy compared to previous studies using the WESAD dataset.