• 한국인터넷방송통신학회논문지
• /
• 제12권1호
• /
• pp.17-27
• /
• 2012

방송 기반의 데이터 전파는 모바일 컴퓨팅 환경에서 보편적으로 받아들여지는 통신 방식이다. 그러나 많은 데이터를 방송해야 할 경우에 방송 채널의 순차성으로 인하여 원하는 데이터를 수신하기까지의 예상 지연시간이 증가한다. 이 대기시간을 줄이기 위하여 본 논문은 다중 채널에 적절하게 데이터를 할당하기 위한 주제를 연구하여 NODA(near optimal data allocation)로 명명된 새로운 데이터 할당 기법을 제안한다. 제안된 기법은 우선 각 방송 채널이 방송해야 하는 데이터들을 결정하기 위하여 평균 예상지연시간의 이론적 하한 값에 기초하여 전체 데이터들을 K개의 그룹으로 분할한다. 그리고 나서 NODA는 동일 방송 채널에 할당된 데이터들을 다른 빈도로 방송하기 위하여, 각 그룹의 데이터들을 확장된 동적 프로그래밍 알고리즘을 사용하여 B개의 그룹으로 분할한다. 제안된 기법은 동일 채널에 할당된 데이터들을 그들의 인기도를 반영하여 방송할 수 있기 때문에 평균 예상지연시간을 최소화할 수 있다.

• 한국인터넷방송통신학회논문지
• /
• 제11권2호
• /
• pp.63-72
• /
• 2011

방송 기반의 데이터 전파는 모바일 컴퓨팅 환경에서 보편적으로 받아들여지는 통신 방식이다. 그러나 많은 데이터를 방송해야 할 경우에 방송 채널의 순차성으로 인하여 원하는 데이터를 수신하기까지의 예상 지연시간이 증가한다. 이 대기시간을 줄이기 위하여 본 논문은 다중 채널을 통하여 데이터를 방송하는 문제를 연구한다. 기존의 기법들은 각 채널의 평형 방송을 가정하고 데이터들을 액세스 확률에 근거하여 분할하고 다중 채널에 할당한다. 만일 동일 채널에 할당된 데이터들도 그들의 액세스 확률에 근거하여 방송빈도를 다르게 한다면 성능은 더 향상될 것이다. 이와 같은 관점에서 본 논문은 동일 채널에 할당된 데이터들의 액세스 확률의 차이를 방송 스케줄에 반영할 수 있는 두 단계 동적프로그래밍 기법으로 명명된 새로운 방송 스케줄링 기법을 제안한다.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.487-491
• /
• 2009

The inverse halftoning is processing to restore the image made binary to former step image. There are smoothing and gaussian filtering in the technique so far. However, there are still a lot of insufficient points in past inverse halftoning. The removal of the noise and the edge enhansment are closely related in inverse halftoning. It is difficult to do both the noise rednctiom and the edge enhansment in high accuracy at the same time in the technique so far. The technique that can achieve both the removal of the noise and the emphasis of Edge at the same time is expected as future tasks. Then, it was tried to apply the Kalmann filtering to inverse halftoning. In the actual experiment, the effectiveness of the application of the Kalmann filtering to inverse halftoning comparing it with the technique so far was shown.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 1998년도 Proceedings of International Workshop on Advanced Image Technology
• /
• pp.122-127
• /
• 1998

In this paper, we focus on the information retrieval of human individuals who are recorded on the video database. Our purpose is to index persons by their faces or voice and to retrieve their existing time sections on the video data. The database system can track as well as extract a face or voice of a certain person and construct a model of the individual person in self-organization mode. If he appears again at different time, the system can put the mark of the same person to the associated frames. In this way, the same person can be retrieved even if the system does not know his exact name. As the face and speaker modeling, a subspace method is employed to improve the indexing accuracy.

• 한국컴퓨터정보학회논문지
• /
• 제24권8호
• /
• pp.51-58
• /
• 2019

In this paper, we propose to characteristics of mobile phone dance live broadcast platform and the second-generation technology acceptance model TAM2, this paper established the user acceptance model of mobile phone dance live broadcast platform, aiming to study the influencing factors of users' acceptance on mobile phone dance live broadcast platform. Based on the empirical analysis of user survey, the model is validated, and the relationship between variables in the model is clarified. It is also confirmed that human-computer interaction, scene interaction, relationship interaction and other factors will affect the user's acceptance on mobile phone dance live broadcast platform. At the same time, based on the relationship among variables obtained in the research, this paper tries to analyze how the variables affect each other based on the actual practice of mobile phone dance live broadcast platform. In addition, the video design strategy and marketing strategy for further development of mobile phone dance live broadcast platform are given to help the platform and dance creators to carry out better promotion on the mobile side. In the end, this paper summarizes the shortcomings of this study and points out further research directions in the future, providing a reference for researchers in the field of mobile phone dance live broadcast platform acceptance.

• 말소리와 음성과학
• /
• 제11권4호
• /
• pp.89-96
• /
• 2019

본 논문에서는 방송 콘텐츠에서 음성 구간 검출을 효과적으로 할 수 있는 심층 학습 모델 구조를 제안한다. 또한 특징 벡터의 시간적 변화를 학습하기 위한 다중 스케일 시간 확장 합성곱 층을 제안한다. 본 논문에서 제안한 모델의 성능을 검증하기 위하여 여러 개의 비교 모델을 구현하고, 프레임 단위의 F-score, precision, recall을 계산하여 보여 준다. 제안 모델과 비교 모델은 모두 같은 학습 데이터로 학습되었으며, 모든 모델은 다양한 장르(드라마, 뉴스, 다큐멘터리 등)로 구성되어 있는 한국 방송데이터 32시간을 이용하여 모델을 학습되었다. 제안 모델은 한국 방송데이터에서 F-score 91.7%로 가장 좋은 성능을 보여주었다. 또한 영국과 스페인 방송 데이터에서도 F-score 87.9%와 92.6%로 가장 높은 성능을 보여주었다. 결과적으로 본 논문의 제안 모델은 특징 벡터의 시간적 변화를 학습하여 음성 구간 검출 성능 향상에 기여할 수 있었다.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 1997년도 Proceedings International Workshop on New Video Media Technology
• /
• pp.75-81
• /
• 1997

A new DTPP (Desk-Top Program Production) system has been developed that enables multiple program producers (directors) working at different locations to collaborate over a computer network and prepare a single program for broadcasting. In this system, information is shared among users by exchanging data edited on non-linear editing terminals in program post-production work over a network in real time. In short, the new DTPP system provides a collaborative work space for producing TV programs. The system does not make use of a special server for collaborative work but rather multiple interconnected editing terminals having the same functions. In this configuration, data at a terminal which has just been edited by some operation is forwarded to all other connected terminals for updating. This form of information sharing, however, requires that some sort of data synchronizing method be established since multiple terminals are operating on the same data simultaneously. We therefore adopt a method whereby the system synchronizes the clocks on each terminal at the time of connection and sends an operation time stamp together with edited data. This enables most recently modified data to be identified and all information on all terminals to be updated appropriately. This paper provides an overview of this new collaborative DTPP system and describes the techniques for exchanging edited data and synchronizing data.

• Journal of Communications and Networks
• /
• 제18권3호
• /
• pp.375-386
• /
• 2016

Most existing studies on broadcast services in orthogonal frequency division multiplexing (OFDM) systems have focused on how to allocate the transmission power to the subcarriers. However, because a broadcasting system must guarantee quality of service to all users, the performance of the broadcast service dominantly depends on the channel state of the user who has the lowest received signal-to-noise ratio among users. To reduce the effect of the worst user on the system performance, we propose a joint delivery scheme of unicast and broadcast transmissions in OFDM systems with broadcast and unicast best-effort users. In the proposed joint delivery scheme, the BS delivers the broadcast information using both the broadcast and unicast subcarriers at the same time in order to improve the performance of the broadcast service. The object of the proposed scheme is to minimize the outage probability of the broadcast service while maximizing the sum-rate of best-effort users. For the proposed joint delivery scheme, we develop an adaptive power and subcarrier allocation algorithm under the constraint of total transmission power. This paper shows that the optimal power allocation over each subcarrier in the proposed scheme has a multi-level water filling form. Because the power allocation and the subcarrier assignment problems should be jointly solved, we develop an iterative algorithm to find the optimal solution. Numerical results show that the proposed joint delivery scheme with adaptive power and subcarrier allocation outperforms the conventional scheme in terms of the outage probability of the broadcast service and the sum-rate of best-effort users.

• ETRI Journal
• /
• 제38권3호
• /
• pp.435-443
• /
• 2016

In this paper, an efficient beam tracking algorithm for a regularized zero-forcing (RZF) approach in slowly fading multiple-input and single-output (MISO) broadcast channels is considered. By modifying an RZF equation, an RZF beam tracking algorithm is proposed using matrix perturbation theory. The proposed algorithm utilizes both beams from the previous time step and channel difference (between the previous and current time steps) to calculate the RZF beams. The tracking performance of the proposed algorithm is analyzed in terms of the mean square error (MSE) between a tracking approach and an exact recomputing approach, and in terms of the additional MSE caused by the beam tracking error at the receiver. Numerical results show that the proposed algorithm has almost the same performance as the exact recomputing approach in terms of the sum rate.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2020년도 추계학술대회
• /
• pp.244-247
• /
• 2020

Data augmentation has been helpful in improving the performance in deep learning, when we have a limited data and random erasing is one of the augmentations that have shown impressive performance in deep learning in multiple domains. But the main issue is that sometime it loses good features when randomly selected region is erased by some random values, that does not improve performance as it should. We target that problem in way that good features should not be lost and also want random erasing at the same time. For that purpose, we introduce new augmentation technique named Intra-Class Random Erasing (ICRE) that focuses on data to learn robust features of the same class samples by randomly exchanging randomly selected region. We perform multiple experiments by using different models including resnet18, VGG16 over variety of the datasets including ESC10, UrbanSound8K. Our approach has shown effectiveness over others methods including random erasing.