• 한국멀티미디어학회논문지
• /
• 제23권12호
• /
• pp.1464-1475
• /
• 2020

Due to the spread of COVID-19, the online lecture has become more prevalent. However, it was found that a lot of students and professors are experiencing lack of communication. This study is therefore designed to improve interactive communication between professors and students in real-time online lectures. To do so, we explore deep learning approaches for automatic recognition of students' facial expressions and classification of their understanding into 3 classes (Understand / Neutral / Not Understand). We use 'BlazeFace' model for face detection and 'ResNet-GRU' model for facial expression recognition (FER). We name this entire process 'Degree of Understanding (DoU)' algorithm. DoU algorithm can analyze a multitude of students collectively and present the result in visualized statistics. To our knowledge, this study has great significance in that this is the first study offers the statistics of understanding in lectures using FER. As a result, the algorithm achieved rapid speed of 0.098sec/frame with high accuracy of 94.3% in CPU environment, demonstrating the potential to be applied to real-time online lectures. DoU Algorithm can be extended to various fields where facial expressions play important roles in communications such as interactions with hearing impaired people.

• Fisheries and Aquatic Sciences
• /
• 제25권3호
• /
• pp.151-157
• /
• 2022

Various approaches have been applied to transform aquaculture from a manual, labour-intensive industry to one dependent on automation technologies in the era of the fourth industrial revolution. Technologies associated with the monitoring of physical condition have successfully been applied in most aquafarm facilities; however, real-time biological monitoring systems that can observe fish condition and behaviour are still required. In this study, we used a video recorder placed on top of a fish tank to observe the swimming patterns of rock bream (Oplegnathus fasciatus), first one fish alone and then a group of five fish. Rock bream in the video samples were successfully identified using the you-only-look-once v3 algorithm, which is based on the Darknet-53 convolutional neural network. In addition to recordings of swimming behaviour under normal conditions, the swimming patterns of fish under abnormal conditions were recorded on adding an anaesthetic or lowering the salinity. The abnormal conditions led to changes in the velocity of movement (3.8 ± 0.6 cm/s) involving an initial rapid increase in speed (up to 16.5 ± 3.0 cm/s, upon 2-phenoxyethanol treatment) before the fish stopped moving, as well as changing from swimming upright to dying lying on their sides. Machine learning was applied to datasets consisting of normal or abnormal behaviour patterns, to evaluate the fish behaviour. The proposed algorithm showed a high accuracy (98.1%) in discriminating normal and abnormal rock bream behaviour. We conclude that artificial intelligence-based detection of abnormal behaviour can be applied to develop an automatic bio-management system for use in the aquaculture industry.

• Smart Structures and Systems
• /
• 제31권4호
• /
• pp.365-381
• /
• 2023

The existing vision-based techniques for inspection and condition assessment of civil infrastructure are mostly manual and consequently time-consuming, expensive, subjective, and risky. As a viable alternative, researchers in the past resorted to deep learning-based autonomous damage detection algorithms for expedited post-disaster reconnaissance of structures. Although a number of automatic damage detection algorithms have been proposed, the scarcity of labeled training data remains a major concern. To address this issue, this study proposed a semi-supervised learning (SSL) framework based on consistency regularization and cross-supervision. Image data from post-earthquake reconnaissance, that contains cracks, spalling, and exposed rebars are used to evaluate the proposed solution. Experiments are carried out under different data partition protocols, and it is shown that the proposed SSL method can make use of unlabeled images to enhance the segmentation performance when limited amount of ground truth labels are provided. This study also proposes DeepLab-AASPP and modified versions of U-Net++ based on channel-wise attention mechanism to better segment the components and damage areas from images of reinforced concrete buildings. The channel-wise attention mechanism can effectively improve the performance of the network by dynamically scaling the feature maps so that the networks can focus on more informative feature maps in the concatenation layer. The proposed DeepLab-AASPP achieves the best performance on component segmentation and damage state segmentation tasks with mIoU scores of 0.9850 and 0.7032, respectively. For crack, spalling, and rebar segmentation tasks, modified U-Net++ obtains the best performance with Igou scores (excluding the background pixels) of 0.5449, 0.9375, and 0.5018, respectively. The proposed architectures win the second place in IC-SHM2021 competition in all five tasks of Project 2.

• 융합신호처리학회논문지
• /
• 제23권2호
• /
• pp.62-69
• /
• 2022

의료현장에서는 최근 디지털 헬스케어의 중요성이 대두되면서, 다양한 형태의 생체신호 측정 관련 연구가 활발히 진행되고 있다. 생체신호 중 가장 중요한 신호로 심전도를 들 수 있으며, 특히 부정맥 환자에 있어 심전도 신호의 연속 모니터링은 매우 중요하다. 부정맥은 동결절(sinus node), 동빈맥(sinus tachycardia), 심방조기수축(atrial premature beat, APB), 심실세동 (ventricular fibrillation) 등으로 그 발병원에 따른 형태가 다양하며, 발병 이후의 예후가 좋지 않으므로 일상 중 연속 모니터링은 부정맥의 조기 진단과 치료방향 설정에서 매우 중요하다. 부정맥 환자의 심전도 신호는 매우 불안정하며, 부정맥을 자동 검출하기 위한 주요 특징점으로 작용하는 정확한 R-peak 포인트의 검출이 어렵다. 본 연구에서는 연속 측정하는 홀터 심전도 모니터링 기기와 분석용 소프트웨어를 개발하였으며, 부정맥 데이터베이스를 통해 심전도 신호의 R-peak 효용성을 확인하였다. 향후 연구에서는 다양한 발병원인으로 인한 부정맥의 형태적 구분 및 예측을 위한 알고리즘과 임상 데이터에 근거한 유효성 검증에 관한 추가 연구가 필요하다.

• 품질경영학회지
• /
• 제51권2호
• /
• pp.283-296
• /
• 2023

Purpose: The purpose of this study is to actually implement and verify whether welding defects can be detected in real time by utilizing deep learning AI solutions in the welding process of electric vehicle hairpin winding motors. Methods: AI's function and technological elements using synthetic neural network were applied to existing electric vehicle hairpin winding motor laser welding process by making special hardware for detecting electric vehicle hairpin motor laser welding defect. Results: As a result of the test applied to the welding process of the electric vehicle hairpin winding motor, it was confirmed that defects in the welding part were detected in real time. The accuracy of detection of welds was achieved at 0.99 based on mAP@95, and the accuracy of detection of defective parts was 1.18 based on FB-Score 1.5, which fell short of the target, so it will be supplemented by introducing additional lighting and camera settings and enhancement techniques in the future. Conclusion: This study is significant in that it improves the welding quality of hairpin winding motors of electric vehicles by applying domestic artificial intelligence solutions to laser welding operations of hairpin winding motors of electric vehicles. Defects of a manufacturing line can be corrected immediately through automatic welding inspection after laser welding of an electric vehicle hairpin winding motor, thus reducing waste throughput caused by welding failure in the final stage, reducing input costs and increasing product production.

• 한국지형학회지
• /
• 제26권2호
• /
• pp.69-84
• /
• 2019

The purpose of this study is to extract lineaments in the southeastern part of the Gaeum Fault System, and to understand their characteristics and a relationship between them and fault activation. The lineaments were extracted using a multi-layered analysis based on a digital elevation model (5 m resolution), aerial photos, and satellite images. First-grade lineaments inferred as an high-activity along them were classified based on the displacement of the Quaternary deposits and the distribution of fault-related landforms. The results of classifying the first-grade lineaments were verified by fieldwork and electrical resistivity survey. In the study area of 510 km², a total of 222 lineaments was identified, and their total length was 333.4 km. Six grade lineaments were identified, and their total length was 11.2 km. The lineaments showed high-density distribution in the region along the Geumcheon, Gaeum, Ubo fault, and a boundary of the Hwasan cauldron consisting the Gaeum Fault System. They generally have WNW-ESE trend, which is the same direction with the strike of Gaeum Fault System. Electrical resistivity survey was conducted on eight survey lines crossing the first-grade lineament. A low-resistivity zone, which is assumed to be a fault damage zone, has been identified across almost all survey lines (except for only one survey line). The visual (naked eyes) detecting of the lineament was evaluated to be less objectivity than the automatic extraction using the algorithm. However, the results of electrical resistivity survey showed that first-grade lineament extracted by visual detecting was 83% reliable for inferred fault detection. These results showed that objective visual detection results can be derived from multi-layered analysis based on tectonic geomorphology.

• 대한의용생체공학회:의공학회지
• /
• 제44권5호
• /
• pp.346-353
• /
• 2023

In this paper, we propose an innovative approach that leverages deep learning to find optimal reference points for achieving precise tooth segmentation in three-dimensional tooth point cloud data. A dataset consisting of 350 aligned maxillary and mandibular cloud data was used as input, and both end coordinates of individual teeth were used as correct answers. A two-dimensional image was created by projecting the rendered point cloud data along the Z-axis, where an image of individual teeth was created using an object detection algorithm. The proposed algorithm is designed by adding various modules to the Unet model that allow effective learning of a narrow range, and detects both end points of the tooth using the generated tooth image. In the evaluation using DSC, Euclid distance, and MAE as indicators, we achieved superior performance compared to other Unet-based models. In future research, we will develop an algorithm to find the reference point of the point cloud by back-projecting the reference point detected in the image in three dimensions, and based on this, we will develop an algorithm to divide the teeth individually in the point cloud through image processing techniques.

• 대한원격탐사학회지
• /
• 제39권5_1호
• /
• pp.637-654
• /
• 2023

최근 지구관측 위성이 급격히 발전함에 따라 사용자의 수가 증가하고 있다. 이에 따라 지구관측위성위원회(Committee on Earth Observation Satellites, CEOS)에서는 분석준비자료(Analysis Ready Data, ARD)라는 개념을 제안하고 분석준비자료의 요구 조건을 CEOS ARD for Land (CARD4L)로 정의하여 사용자 친화적인 위성영상을 제공하기 위해 노력하고 있다. 분석준비자료에는 육상분석에 불필요한 픽셀이 식별된 마스크(Unusable Data Mask, UDM)가 영상과 함께 제공되어야 한다. UDM의 종류는 구름, 구름 그림자, 지형그림자 등이 있다. 지형그림자는 지형기복이 큰 산악지형에서 발생되며 지형그림자가 생긴 지역은 복사조도가 낮기 때문에 분석 결과에 오류를 야기시킨다. 기존 지형그림자 탐지연구는 지형그림자 보정을 위해 지형그림자 픽셀을 탐지하는데 목적을 두었지만, 이것은 지형보정 기법으로 대체 가능하다. 따라서 지형그림자 탐지 목적을 확장할 필요가 있다. 산림과 농업분석을 목적으로 한 차세대중형위성 4호(CAS500-4)의 활용을 위해 본 연구에서는 지형그림자 탐지 범위를 태양의 영향을 적게 받는 지역까지 확장하였다. 본 논문은 남북한을 대상으로 지형그림자 마스크 생성을 위해 지형그림자 탐지 가능성을 분석하는데 목적이 있다. 지형그림자 탐지를 위해서 태양의 위치, 지표면의 경사와 경사방향을 이용한 음영기복 알고리즘을 사용하였다. 한반도를 촬영한 5 m급 공간해상도의 RapidEye 영상과 10 m급 공간해상도의 Sentinel-2 영상들을 대상으로 참값과 비교하며 최적의 음영기복 임계값을 결정하였다. 결정된 임계값을 사용하여 지형 그림자 탐지를 수행하고 결과를 분석하였다. 정성적 결과로는 전체적으로 참값과의 형상이 유사함을 확인하였다. 정량적 실험결과는 F1 score가 대부분 0.8에서 0.94 사이인 것을 확인하였다. 본 연구 결과를 바탕으로 남북한을 대상으로 자동적인 지형그림자 탐지가 잘 수행됨을 확인하였다.

• 융합신호처리학회논문지
• /
• 제7권3호
• /
• pp.116-121
• /
• 2006

본 논문에서는 다양한 잡음환경에서의 실시간 적응화 기법을 적용하기 위한 선결 과제로 다차원 음성 특정 벡터를 저차원으로 축소하는 방법을 제안한다. 제안된 방법은 특징 벡터를 확률 우도 값으로 매핑시켜 비선형적으로 축소하는 방법으로 음성 / 비음성의 분류는 우도비 검증 (Likelihood Ratio Test; LRT) 을 이용하여 분류하였다. 실험 결과 고차원 특징 벡터를 이용하여 분류한 결과와 대등하게 분류됨을 확인할 수 있었다. 그리고, 제안된 방법에 의해 검출된 음성 데이터를 이용한 음성인식 실험에서도 10차 MFCC(Mel-Frequency Cepstral Coefficient)를 사용하여 분류한 경우와 대등한 인식률을 보여주었다.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2011년도 춘계학술대회 논문집
• /
• pp.1009-1013
• /
• 2011

Seoul Metro Line 1 to 4 guests per day to 4.5 million people have been using the subway. This is close to the city population, 57 percent of people use the subway and found that 57 percent of people use the subway and found that this is close to the city population. Motor vehicle indoor air quality, especially of the major factors affecting is passenger's clothing, cosmetics, adhesives and formaldehyde by the action and so are able to keep. Enclosure 30ppm formaldehyde during prolonged exposure at concentrations above the nose, bronchial cough and a burning can cause symptoms. It is necessary to introduce an appropriate ventilation system. 1-4 Line Press in 2008, leaving the subway, and normally the train measured in room air quality. Measurements in 2005, $19.3{\sim}83{\mu}g/m^3$, 2008 Year ND ~ $61.7{\mu}g/m^3$ is. When congestion(rush hour) to temporarily increase the formaldehyde can result in a higher number. Automatic detection of formaldehyde and improve ventilation to a practical system, and it is necessary to chatneun. In research since 2006, Removal of formaldehyde were analyzed for the study. the passengers on the effects of formaldehyde in rush hour, the plan for increasing the ventilation through the analysis of various factors, such as electric vehicle practical ways to improve air quality have been studied.