• Annual Conference of KIPS
• /
• 2023.11a
• /
• pp.11-13
• /
• 2023

의미론적 분할(Semantic Segmentation)은 이미지 내의 객체 및 배경을 픽셀 단위로 분류하는 작업으로 정밀한 탐지가 요구되는 분야에서 활발히 연구되고 있다. 기존 어텐션 기법은 의미론적 분할의 다운샘플링(Downsampling) 과정에서 발생하는 정보손실을 완화하기 위해 널리 사용됐지만 고정된 Convolution 필터의 형태 때문에 객체의 형태에 따라 유동적으로 대응하지 못했다. 본 논문에서는 이를 보완하고자 Deformable Convolution과 셀프어텐션(Self-attention) 구조기반 어텐션 모듈을 사용한 의미론적 분할 모델을 제안한다.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.10
• /
• pp.1-9
• /
• 2022

In this paper, we propose the Deep Learning-Based Companion Animal Abnormal Behavior Detection Service, which using video and sensor data. Due to the recent increase in households with companion animals, the pet tech industry with artificial intelligence is growing in the existing food and medical-oriented companion animal market. In this study, companion animal behavior was classified and abnormal behavior was detected based on a deep learning model using various data for health management of companion animals through artificial intelligence. Video data and sensor data of companion animals are collected using CCTV and the manufactured pet wearable device, and used as input data for the model. Image data was processed by combining the YOLO(You Only Look Once) model and DeepLabCut for extracting joint coordinates to detect companion animal objects for behavior classification. Also, in order to process sensor data, GAT(Graph Attention Network), which can identify the correlation and characteristics of each sensor, was used.

• The KIPS Transactions:PartC
• /
• v.15C no.5
• /
• pp.351-358
• /
• 2008

Recently, as network flooding attacks such as DoS/DDoS and Internet Worm have posed devastating threats to network services, rapid detection and proper response mechanisms are the major concern for secure and reliable network services. However, most of the current Intrusion Detection Systems(IDSs) focus on detail analysis of packet data, which results in late detection and a high system burden to cope with high-speed network environment. In this paper we propose a lightweight and fast detection mechanism for traffic flooding attacks. Firstly, we use SNMP MIB statistical data gathered from SNMP agents, instead of raw packet data from network links. Secondly, we use a machine learning approach based on a Support Vector Machine(SVM) for attack classification. Using MIB and SVM, we achieved fast detection with high accuracy, the minimization of the system burden, and extendibility for system deployment. The proposed mechanism is constructed in a hierarchical structure, which first distinguishes attack traffic from normal traffic and then determines the type of attacks in detail. Using MIB data sets collected from real experiments involving a DDoS attack, we validate the possibility of our approaches. It is shown that network attacks are detected with high efficiency, and classified with low false alarms.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.469-472
• /
• 2022

With the recent development of intelligent transportation systems, various technologies applying deep learning technology are being used. To crackdown on illegal vehicles and criminal vehicles driving on the road, a vehicle type classification system capable of accurately determining the type of vehicle is required. This study proposes a vehicle type classification system optimized for mobile traffic control systems using YOLO(You Only Look Once). The system uses a one-stage object detection algorithm YOLOv5 to detect vehicles into six classes: passenger cars, subcompact, compact, and midsize vans, full-size vans, trucks, motorcycles, special vehicles, and construction machinery. About 5,000 pieces of domestic vehicle image data built by the Korea Institute of Science and Technology for the development of artificial intelligence technology were used as learning data. It proposes a lane designation control system that applies a vehicle type classification algorithm capable of recognizing both front and side angles with one camera.

• Journal of Cadastre & Land InformatiX
• /
• v.52 no.1
• /
• pp.81-93
• /
• 2022

Recently, there have been many studies on object recognition around the vehicle and recognition of traffic signs and traffic lights in autonomous driving. In particular, such the recognition of traffic lights is one of the core technologies in autonomous driving. Therefore, many studies for such the recognition of traffic lights have been performed, the studies based on various deep learning models have increased significantly in recent. In addition, as a high-quality AI training data set for voice, vision, and autonomous driving is released on AIHub, it makes it possible to develop a recognition model for traffic lights suitable for the domestic environment using the data set. In this study, we developed a recognition model for traffic lights that can be used in Korea using the AIHub's training data set. In particular, in order to improve the recognition performance, we used various models of YOLOv4 and YOLOv5, and performed our recognition experiments by defining various classes for the training data. In conclusion, we could see that YOLOv5 shows better performance in the recognition than YOLOv4 and could confirm the reason from the architecture comparison of the two models.

• Smart Media Journal
• /
• v.12 no.10
• /
• pp.71-84
• /
• 2023

Paralichthys olivaceus accounts for a large proportion, accounting for more than half of Korea's aquaculture industry. However, about 25-30% of the total breeding volume throughout the year occurs due to diseases, which has a very bad impact on the economic feasibility of fish farms. For the economic growth of Paralichthys olivaceus farms, it is necessary to quickly and accurately diagnose disease symptoms by automating the diagnosis of Paralichthys olivaceus diseases. In this study, we create training data using innovative data collection methods, refining data algorithms, and techniques for partitioning dataset, and compare the Paralichthys olivaceus disease symptom detection performance of four object detection deep learning models(such as YOLOv8, Swin, Vitdet, MvitV2). The experimental findings indicate that the YOLOv8 model demonstrates superiority in terms of average detection rate (mAP) and Estimated Time of Arrival (ETA). If the performance of the AI model proposed in this study is verified, Paralichthys olivaceus farms can diagnose disease symptoms in real time, and it is expected that the productivity of the farm will be greatly improved by rapid preventive measures according to the diagnosis results.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_3
• /
• pp.1777-1788
• /
• 2022

The reservoir is a major water supply source in the domestic agricultural environment, and the monitoring of water storage of reservoirs is important for the utilization and management of agricultural water resource. Remote sensing via satellite imagery can be an effective method for regular monitoring of widely distributed objects such as reservoirs, and in this study, image classification and image segmentation algorithms are applied to Sentinel-1 Synthetic Aperture Radar (SAR) imagery for water body detection in 53 reservoirs in South Korea. Six algorithms are used: Neural Network (NN), Support Vector Machine (SVM), Random Forest (RF), Otsu, Watershed (WS), and Chan-Vese (CV), and the results of water body detection are evaluated with in-situ images taken by drones. The correlations between the in-situ water surface area and detected water surface area from each algorithm are NN 0.9941, SVM 0.9942, RF 0.9940, Otsu 0.9922, WS 0.9709, and CV 0.9736, and the larger the scale of reservoir, the higher the linear correlation was. WS showed low recall due to the undetected water bodies, and NN, SVM, and RF showed low precision due to over-detection. For water body detection through SAR imagery, we found that aquatic plants and artificial structures can be the error factors causing undetection of water body.

• The Journal of Bigdata
• /
• v.9 no.1
• /
• pp.89-100
• /
• 2024

In the field of computer vision, models such as You Look Only Once (YOLO) and ResNet are widely used due to their real-time performance and high accuracy. However, to apply these models in real-world environments, factors such as runtime compatibility, memory usage, computing resources, and real-time conditions must be considered. This study compares the characteristics of three deep model runtimes: ONNX Runtime, TensorRT, and OpenCV DNN, and analyzes their performance on two models. The aim of this paper is to provide criteria for runtime selection for practical applications. The experiments compare runtimes based on the evaluation metrics of time, memory usage, and accuracy for vehicle license plate recognition and classification tasks. The experimental results show that ONNX Runtime excels in complex object detection performance, OpenCV DNN is suitable for environments with limited memory, and TensorRT offers superior execution speed for complex models.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.42 no.1
• /
• pp.117-126
• /
• 2022

Currently, the number of bridge structures in Korea is continuously increasing and enlarged, and the number of old bridges that have been in service for more than 30 years is also steadily increasing. Bridge aging is being treated as a serious social problem not only in Korea but also around the world, and the existing manpower-centered inspection method is revealing its limitations. Recently, various bridge damage detection studies using deep learning-based image processing algorithms have been conducted, but due to the limitations of the bridge damage data set, most of the bridge damage detection studies are mainly limited to one type of crack, which is also based on a close set classification model. As a detection method, when applied to an actual bridge image, a serious misrecognition problem may occur due to input images of an unknown class such as a background or other objects. In this study, five types of bridge damage including crack were defined and a data set was built, trained as a deep learning model, and an open set recognition-based bridge multiple damage recognition model applied with OpenMax algorithm was constructed. And after performing classification and recognition performance evaluation on the open set including untrained images, the results were analyzed.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.11
• /
• pp.305-313
• /
• 2019

Recently, spatial information is being constructed actively based on the images obtained by drones. Because occlusion areas occur due to buildings as well as many obstacles, such as trees, pedestrians, and banners in the urban areas, an efficient way to resolve the problem is necessary. Instead of the traditional way, which replaces the occlusion area with other images obtained at different positions, various models based on deep learning were examined and compared. A comparison of a type of feature descriptor, HOG, to the machine learning-based SVM, deep learning-based DNN, CNN, and RNN showed that the CNN is used broadly to detect and classify objects. Until now, many studies have focused on the development and application of models so that it is impossible to select an optimal model. On the other hand, the upgrade of a deep learning-based detection and classification technique is expected because many researchers have attempted to upgrade the accuracy of the model as well as reduce the computation time. In that case, the procedures for generating spatial information will be changed to detect the occlusion area and replace it with simulated images automatically, and the efficiency of time, cost, and workforce will also be improved.