• Annual Conference of KIPS
• /
• 2020.11a
• /
• pp.556-559
• /
• 2020

최근 태풍, 지진, 산불, 산사태, 전쟁 등 다양한 재난 상황으로 인한 인명피해와 자금 손실이 꾸준히 발생하고 있고 현재 이를 예방하고 복구하기 위해 많은 인력과 자금이 소요되고 있는 실정이다. 이러한 여러 재난 상황을 미리 감시하고 재난 발생의 빠른 인지 및 대처를 위해 본 논문에서는 인공지능 기반의 재난 드론 시스템을 설계 및 개발하였다. 본 연구에서는 사람이 감시하기 힘든 지역에 여러 대의 재난 드론을 이용하며 딥러닝 기반의 최단 경로 알고리즘을 적용해 각각의 드론이 최적의 경로로 효율적 탐색을 실시한다. 또한 드론의 근본적 문제인 배터리 용량 부족에 대한 문제점을 해결하기 위해 Ant Colony Optimization (ACO) 기술을 이용하여 각 드론의 최적 경로를 결정하게 된다. 제안한 시스템 구현을 위해 여러 재난 상황 중 산불 상황에 적용하였으며 전송된 데이터를 기반으로 산불지도를 만들고, 빔프로젝터를 탑재한 드론이 출동한 소방관에게 산불지도를 시각적으로 보여주었다. 제안한 시스템에서는 여러 대의 드론이 최적 경로 탐색 및 객체인식을 동시에 수행함으로써 빠른 시간 내에 재난 상황을 인지할 수 있다. 본 연구를 바탕으로 재난 드론 인프라를 구축하고 조난자 탐색(바다, 산, 밀림), 드론을 이용한 자체적인 화재진압, 방범 드론 등에 활용할 수 있다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.483-483
• /
• 2022

댐과 저수지의 운영 최적화를 위한 수문학적 예보는 현재 수동적인 댐 운영이 주를 이루면서 활용도가 높지 않다. 불확실한 기후변화나 기후재난 상황에서 우리 사회에 악영향을 최소화하기 위해 선제적으로 대응/대비할 수 있는 댐 운영 방안이 불가피하다. 강우량 예측 기술은 기후변화로 인해 제한적인 상황이다. 실례로, 2020년 8월에 섬진강의 댐이 극심한 집중 강우로 인해 무너지는 사태가 발생하였고 이로 인해 지역사회에 막대한 경제적 피해가 발생하였다. 선제적 댐 방류량 운영 기술은 또한 환경적인 변화로 인한 영향을 완화하기 위해 필요한 것이다. 제한적인 기상 예보 기술을 극복하고자 심화학습이나 강화학습 같은 인공지능 모델들의 활용성에 대한 연구가 시도되고 있다. 따라서 본 연구는 섬진강 댐의 시간당 수문 데이터를 이용하여 댐 운영을 위한 심화학습 모델을 개발하고 그 활용도를 평가하였다. 댐 운영을 위한 심화학습 모델로서 시계열 데이터 예측에 적합한 Long Sort Term Memory(LSTM)과 Gated Recurrent Unit(GRU) 알고리즘을 구축하고 댐 수위를 예측하였다. 분석 자료는 WAMIS에서 제공하는 2000년부터 2021년까지의 시간당 데이터를 사용하였다. 입력 데이터로서 시간당 유입량, 강우량과 방류량을, 출력 데이터로서 시간당 수위 자료를 각각 사용하였으며. 결정계수(R2 Score)를 통해 모델의 예측 성능을 평가하였다. 댐 수위 예측값 개선을 위해 하이퍼파라미터의 '최적값'이 존재하는 범위를 줄여나가는 하이퍼파라미터 최적화를 두 가지 방법으로 진행하였다. 첫 번째 방법은 수동적 탐색(Manual Search) 방법으로 Sequence Length를 24, 48, 72시간, Hidden Layer를 1, 3, 5개로 설정하여 하이퍼파라미터의 조합에 따른 LSTM와 GRU의 민감도를 평가하였다. 두 번째 방법은 Grid Search로 최적의 하이퍼파라미터를 찾았다. 이 두가지 방법에서는 같은 하이퍼파라미터 안에서 GRU가 LSTM에 비해 더 높은 예측 정확도를 보였고 Sequence Length가 높을수록 정확도가 높아지는 경향을 보였다. Manual Search 방법의 경우 R2가 최대 0.72의 정확도를 보였고 Grid Search 방법의 경우 R2가 0.79의 정확도를 보였다. 본 연구 결과는 가뭄과 홍수와 같은 물 재해에 사전 대응하고 기후변화에 적응할 수 있는 댐 운영 개선에 도움을 줄 수 있을 것으로 판단된다.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.23 no.1
• /
• pp.182-196
• /
• 2024

This study introduces driving pattern analysis and change detection methods using smartphone sensors, based on contrastive learning. These methods characterize driving patterns without labeled data, allowing accurate classification with minimal labeling. In addition, they are robust to domain changes, such as different vehicle types. The study also examined the applicability of these methods to smartphones by comparing them with six lightweight deep-learning models. This comparison supported the development of smartphone-based driving pattern analysis and assistance systems, utilizing smartphone sensors and contrastive learning to enhance driving safety and efficiency while reducing the need for extensive labeled data. This research offers a promising avenue for addressing contemporary transportation challenges and advancing intelligent transportation systems.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.59-64
• /
• 2024

Recently, research on the convergence of drones and artificial intelligence technologies have been conducted in various industrial fields. In this paper, we propose an illegal parking vehicle recognition model using deep learning-based object recognition and classification algorithms. The model of object recognition and classification consist of YOLOv8 and ResNet18, respectively. The proposed model was trained using image data collected in general road environment, and the trained model showed high accuracy in determining illegal parking. From simulation results, it was confirmed that the proposed model has generalization performance to identify illegal parking vehicles from various images.

• Journal of the Korean Society of Radiology
• /
• v.81 no.6
• /
• pp.1305-1333
• /
• 2020

Deep learning has recently achieved remarkable results in the field of medical imaging. However, as a deep learning network becomes deeper to improve its performance, it becomes more difficult to interpret the processes within. This can especially be a critical problem in medical fields where diagnostic decisions are directly related to a patient's survival. In order to solve this, explainable artificial intelligence techniques are being widely studied, and an attention mechanism was developed as part of this approach. In this paper, attention techniques are divided into two types: post hoc attention, which aims to analyze a network that has already been trained, and trainable attention, which further improves network performance. Detailed comparisons of each method, examples of applications in medical imaging, and future perspectives will be covered.

• Journal of Internet Computing and Services
• /
• v.25 no.3
• /
• pp.19-25
• /
• 2024

The worldwide aging population trend is causing an increase in the incidence of major retinal diseases that can lead to blindness, including glaucoma, cataract, and macular degeneration. In the field of ophthalmology, there is a focused interest in diagnosing diseases that are difficult to prevent in order to reduce the rate of blindness. This study proposes a deep learning approach to accurately diagnose ocular diseases in fundus photographs using less data than traditional methods. For this, Convolutional Neural Network (CNN) models capable of effective learning with limited data were selected to classify Conventional Fundus Images (CFI) from various ocular disease patients. The chosen CNN models demonstrated exceptional performance, achieving high Accuracy, Precision, Recall, and F1-score values. This approach reduces manual analysis by ophthalmologists, shortens consultation times, and provides consistent diagnostic results, making it an efficient and accurate diagnostic tool in the medical field.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.5
• /
• pp.61-67
• /
• 2024

In recent years, artificial intelligence and deep learning technologies have made significant advances, and the BERT model has been recognized for its excellent contextual understanding in natural language processing based on the transformer architecture. This performance has the potential to take traditional recommendation systems to the next level. In this study, we adopt an approach that combines a collaborative filtering approach with a deep learning model to improve the performance of recommendation systems. Specifically, we implemented a system that uses BERT to analyze the sentiment of user reviews and embed users based on these review sentiments to find and recommend users with similar tastes. In the process, we also utilized Elasticsearch, an open-source search engine, for quick search and retrieval of recommended results. The approach of analyzing users' textual data to increase the accuracy and personalization of recommendations will play an important role in improving the user experience on various online services in the future.

• Tunnel and Underground Space
• /
• v.31 no.5
• /
• pp.374-384
• /
• 2021

In this study, we propose a new approach for automatic fracture detection in CT scan images of rock specimens. This approach is built on top of two-stage object detection deep learning algorithm called Faster R-CNN with a major modification of using rotated bounding box. The use of rotated bounding box plays a key role in the future work to overcome several inherent difficulties of fracture segmentation relating to the heterogeneity of uninterested background (i.e., minerals) and the variation in size and shape of fracture. Comparing to the commonly used bounding box (i.e., axis-align bounding box), rotated bounding box shows a greater adaptability to fit with the elongated shape of fracture, such that minimizing the ratio of background within the bounding box. Besides, an additional benefit of rotated bounding box is that it can provide relative information on the orientation and length of fracture without the further segmentation and measurement step. To validate the applicability of the proposed approach, we train and test our approach with a number of CT image sets of fractured granite specimens with highly heterogeneous background and other rocks such as sandstone and shale. The result demonstrates that our approach can lead to the encouraging results on fracture detection with the mean average precision (mAP) up to 0.89 and also outperform the conventional approach in terms of background-to-object ratio within the bounding box.

• Journal of the Korean Institute of Landscape Architecture
• /
• v.51 no.3
• /
• pp.166-178
• /
• 2023

This study aims to introduce and assess CNN Deep Learning methods to analyze visual landscape images on social media with embedded user perceptions and experiences. This study analyzed visual landscape images by focusing on a healing place. For the study, seven adjectives related to healing were selected through text mining and consideration of previous studies. Subsequently, 50 evaluators were recruited to build a Deep Learning image. Evaluators were asked to collect three images most suitable for 'healing', 'healing landscape', and 'healing place' on portal sites. The collected images were refined and a data augmentation process was applied to build a CNN model. After that, 15,097 images of 'healing' and 'healing landscape' on portal sites were collected and classified to analyze the visual landscape of a healing place. As a result of the study, 'quiet' was the highest in the category except 'other' and 'indoor' with 2,093 (22%), followed by 'open', 'joyful', 'comfortable', 'clean', 'natural', and 'beautiful'. It was found through research that CNN Deep Learning is an analysis method that can derive results from visual landscape image analysis. It also suggested that it is one way to supplement the existing visual landscape analysis method, and suggests in-depth and diverse visual landscape analysis in the future by establishing a landscape image learning dataset.

• Journal of the Korean Society of Radiology
• /
• v.16 no.5
• /
• pp.661-672
• /
• 2022

After the outbreak of the SARS-CoV2 virus that causes COVID-19, it spreads around the world with the number of infections and deaths rising rapidly caused a shortage of medical resources. As a way to solve this problem, chest X-ray diagnosis using Artificial Intelligence(AI) received attention as a primary diagnostic method. The purpose of this study is to comprehensively analyze the detection of COVID-19 via AI. To achieve this purpose, 292 studies were collected through a series of Classification methods. Based on these data, performance measurement information including Accuracy, Precision, Area Under Cover(AUC), Sensitivity, Specificity, F1-score, Recall, K-fold, Architecture and Class were analyzed. As a result, the average Accuracy, Precision, AUC, Sensitivity and Specificity were achieved as 95.2%, 94.81%, 94.01%, 93.5%, and 93.92%, respectively. Although the performance measurement information on a year-on-year basis gradually increased, furthermore, we conducted a study on the rate of change according to the number of Class and image data, the ratio of use of Architecture and about the K-fold. Currently, diagnosis of COVID-19 using AI has several problems to be used independently, however, it is expected that it will be sufficient to be used as a doctor's assistant.