• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.95-96
• /
• 2021

최근 미디어의 발전으로 빠른 속도로 많은 양의 사람들의 얼굴이 포함된 사진, 동영상들이 인터넷에 업로드 되고 있다. 이러한 현상에 맞춰 인공지능을 활용한 얼굴 인식 기술의 놀라운 발전이 있었으나, 대규모 데이터셋에서 임의의 인물을 검색하는 경우에서는 연산량과 저장공간의 부담이 존재한다. 특히, 인터넷에 존재하는 수많은 불법 촬영물에서 피해자를 정확하고 신속하게 검색하기 위해서는 효율적인 얼굴 검색 시스템이 필요하다. 따라서, 본 논문은 얼굴 특징 추출과 클러스터링을 활용하여 방대한 양의 불법 촬영물 셋에서 피해자 동영상을 효율적으로 검색할 수 있는 기법을 제안한다. 불법 촬영물 동영상 검색 실험 환경을 만들기 위해 YouTube Faces [1] 데이터셋으로 유사 동영상 셋을 만들고 이 환경에서 실험을 진행한다. 얼굴 특징 추출 모델은 ResNet100 네트워크를 CosFace 손실함수와 Glint360K 데이터셋으로 학습시킨 모델 [2]을 사용한다. 추출된 얼굴 특징들을 HAC(Hierarchical Agglomerative Clustering) 알고리즘으로 클러스터링 한 후, 클러스터 대푯값을 통해 얼굴 검색 실험을 했을 때의 실험 결과를 분석한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.3
• /
• pp.361-367
• /
• 2021

Information search in video streaming services such as YouTube is replacing traditional information search services. To find desired detailed information in such a video, users should repeatedly navigate several points in the video, resulting in a waste of time and network traffic. In this paper, we propose a method to assist users in searching for information in a video by using DBSCAN clustering and LSTM. Our LSTM model is trained with a dataset that consists of user search sequences and their final target points categorized by DBSCAN clustering algorithm. Then, our proposed method utilizes the trained model to suggest an expected category for the user's desired target point based on a partial search sequence that can be collected at the beginning of the search. Our experiment results show that the proposed method successfully finds user destination points with 98% accuracy and 7s of the time difference by average.

• The Journal of the Korea Contents Association
• /
• v.13 no.11
• /
• pp.148-158
• /
• 2013

This research attempts to analyse the 'Gangnam style' which has achieved worldwide recognition and became the first YouTube video to reach a billion views on December 2012, since it has released on 15th of July, 2012. Why 'Gangnam style' is so popular and how Cy's music video has spread across the world remarkably in very short period of time? It is suggested that the 'Gangsnam style' phenomenon can be explained by 'Kirtsch' and 'Meme', however there is no pertinent analysis in an academic way. In this respect, this research is concerned with addressing the 'Gangnam style', global cultural phenomenon, by the Kitsch related to the acceptance of global audience and the Meme engaged with global diffusion of music video in a cultural view. Therefore, this research reviews a broad range of literatures and seeks characteristics of two theories, on the ground of this, 'Gangnam style' will be investigated to verify the reason caused cultural phenomenon.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2013.11a
• /
• pp.1483-1486
• /
• 2013

최근 무선통신의 급격한 발전과 스마트 기기의 확산으로 인해 Tving, pooq 등 다양한 모바일 방송 서비스가 급속도로 증가하고 있다. 또한 다양한 영상처리 기법 등이 등장함에 따라 4K, 8K급의 UHD 동영상들이 속속들이 등장하고 있다. 이로 인해 트랜스코딩을 통해 가공되는 동영상의 포맷 및 해상도 또한 매우 다양해질 것으로 전망된다. 현재까지의 트랜스코딩 연구사례는 사용자의 이동환경을 고려한 안정적 QoS 보장 또는 서버의 부하를 줄이기 위한 분산처리 기법 등의 연구 위주로 진행되어 온 것이 현실이다. 하지만 상기 조건(adaptive streaming 및 서버부하 처리)들을 충족시키긴 위해선 보다 효율적인 트랜스코딩 시스템의 제공이 선행 되어야 할 것이다. 이에 따라 본 논문에서는 사용자 관점에서 보다 빨리 스트리밍 서비스를 제공 받기 위하여 우선순위 큐 알고리즘을 적용한 시스템을 설계 및 구현하였다. 검증을 위하여 4가지 콘테이너(.MOV, .FLV, .MKV, .AVI)를 실험대상으로 하였고, 비교 대상 트랜스코딩 시스템은 상용 스트리밍 서비스인 YouTube를 활용하였다. 성능 측정결과, 총 트랜스코딩 완료시간은 YouTube에 비해 41.61%로 시간이 단축되었다. 또한 모바일 TV시청자가 55%를 차지한다는 점을 고려하여 컨트롤 서버에서는 최단시간 서비스 제공을 위하여 저해상도부터 추출하여 스트리밍 서버를 통해 송출하도록 구현하였다. 본 연구결과는 트랜스코딩 성능개선 뿐만 아니라 모바일 대상자를 위한 맞춤형 서비스를 보다 빨리 제공할 수 있을 것이며, 그 수요는 점차 증대될 것으로 예상된다.

• Journal of Digital Convergence
• /
• v.19 no.2
• /
• pp.457-462
• /
• 2021

YouTube has established itself as a major media platform. Businesses give more attention to YouTube's influence as marketing using the platform demonstrates positive effects. Accordingly, this study aims to look into how the quasi-social interaction of the users of beauty content on YouTube affect purchase behavior intentions and provide results as basic data for establishing an efficient marketing strategy. Quasi-social interaction exerted significantly positive influence on purchase behavior and purchase intention. This suggests that the stronger the quasi-social interaction becomes, purchase behavior and purchase intention get stronger. Therefore, strategies for facilitating marketing should be devised by increasing efficient quasi-social interaction according to user characteristics.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.189-192
• /
• 2021

YouTube, the world's largest video sharing platform, is loved by many users in that it provides numerous videos and makes it easy to get helpful information. However, the ratio of like/hate for each video varies according to the subject or upload time, even though they are in the same channel; thus, previous studies try to understand the reason by inspecting some numerical statistics such as the ratio and view count. They can help know how each video is preferred, but there is an explicit limitation to identifying the cause of such preference. Therefore, this study aims to determine the reason that affects the preference through matching between topic words extracted from comments in each video and non-sympathetic types defined in advance. Among the top 10 channels in the field of 'pets' and 'cooking', where outliers occur a lot, the top 10 videos (the threshold of pet: 4.000, the threshold of cooking: 0.723) with the highest ratio were selected. 11,110 comments collected totally, and topics were extracted and matched with non-sympathetic types. The experimental results confirmed that it is possible to predict whether the rate of like/hate would be high or which non-sympathetic type would be by analyzing the comments.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.5
• /
• pp.459-468
• /
• 2022

This study searched and analyzed news and information on the Korea-U.S. Summit to explore which political agenda is spreading among Korean newspapers and social media. The result of the analysis revealed that, on the one hand, the conservative-leaning newspaper, Chosunilbo, covered the unresolved issue between two countries. The principal source of news was the opposition party. On the other hand, the progressive-leaning newspaper, Kyunghany Sinmun, highlighted President Moon's visit to the United States and described the visit to the United States as an achievement. In this paper, the principal source of news is the ruling party. Both conservative and the progressive newspapers did not present a negative view of the United States. In the case of Chosunilbo, it mentioned that foreign policy priority of President Biden is human rights in North Korea. If the two countries do not solve this issue, the relationship between Korea and the United States will not develop further. Second, I searched YouTube videos about the Korea-U.S. summit and conducted a network analysis to understand the influence of YouTube videos and explore their relationship the each other. The results of the analysis revealed that the 10 most influential videos portrayed the Moon government positively. These videos held the achievement of the visit to the United States in highly esteem and framed it positively, similarly to the progressive newspaper.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2013.06a
• /
• pp.176-179
• /
• 2013

본 논문에서는 JCT-VC 에서 2013 년 1 월에 표준화가 완료된 High Efficiency Video Coding (HEVC)과 구글에서 2013 년 6 월에 개발 완료 예정인 VP9 의 압축 효율 비교를 수행한다. HEVC 는 UHD 등 고화질 방송 등에 대응하도록 디자인 되었으며, VP9 은 유튜브 (YouTube) 등과 같은 인터넷 비디오 스트리밍에 적합하도록 디자인되었다. VP9 의 경우 HEVC 와는 달리 로열티 프리 (royalty-free)를 지향하며 오픈소스 (open source) 방식으로 개발이 진행되고 있다. 본 논문에서는 HEVC 와 VP9 의 디자인 차별점을 소개하고, 랜덤 액세스 환경(Random Access, RA)과 저지연 환경 (Low Delay, LD)에서 HEVC 와 VP9 의 압축 효율을 비교한다. 실험 결과에 따르면, 방송 및 패키지 미디어 등에서 많이 사용될 랜덤 액세스 환경에서는 VP9 이 HEVC 대비 32.7% 열세를 보인다. 비디오 컨퍼런스등과 같은 저지연 환경에서는 VP9 이 HEVC 대비 26.7% 열세를 보인다. VP9 의 경우 개발이 완료된 것이 아니므로, 향후 압축 효율의 향상이 있을 것으로 기대된다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2017.05a
• /
• pp.482-484
• /
• 2017

Over the past 70 years, movies and television have revolutionized the way people communicate. However, even with this development, TV has been used only as a means of communication targeting an unspecified number of people due to the restriction of media such as radio waves and movies. However, the development of the Internet and online video has come to a time when 100 million people watch YouTube videos uploaded from the other side of the world by eliminating these restrictions. The message that you want to deliver now can be delivered to anyone, but making the image with these messages remains the last obstacle to communication. To solve these problems, we implemented a web application and a video production program through AWS. This system basically provides the administrator with the video production through the easy interface, the information management and the program on the server on the internet through AWS, the assigned lecture including the computer and the smart phone, the learning materials, And implemented to increase the efficiency of educational video production service.

• Journal of Korea Society of Industrial Information Systems
• /
• v.29 no.1
• /
• pp.91-105
• /
• 2024

Car accidents occur as a result of collisions between vehicles, leading to both vehicle damage and personal and material losses. This study developed a vehicle accident detection model based on 2,550 image frames extracted from car accident videos uploaded to YouTube, captured by CCTV. To preprocess the data, bounding boxes were annotated using roboflow.com, and the dataset was augmented by flipping images at various angles. The You Only Look Once version 8 (YOLOv8) model was employed for training, achieving an average accuracy of 0.954 in accident detection. The proposed model holds practical significance by facilitating prompt alarm transmission in emergency situations. Furthermore, it contributes to the research on developing an effective and efficient mechanism for vehicle accident detection, which can be utilized on devices like smartphones. Future research aims to refine the detection capabilities by integrating additional data including sound.