• Electronics and Telecommunications Trends
• /
• v.36 no.2
• /
• pp.75-82
• /
• 2021

Online action detection (OAD) in a streaming video is an attractive research area that has aroused interest lately. Although most studies for action understanding have considered action recognition in well-trimmed videos and offline temporal action detection in untrimmed videos, online action detection methods are required to monitor action occurrences in streaming videos. OAD predicts action probabilities for a current frame or frame sequence using a fixed-sized video segment, including past and current frames. In this article, we discuss deep learning-based OAD models. In addition, we investigated OAD evaluation methodologies, including benchmark datasets and performance measures, and compared the performances of the presented OAD models.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.40 no.6
• /
• pp.619-625
• /
• 1991

Phenomena of streaming electrification of insulting oil(T.O) is studied where the oil is contacted with solid insulating materials when it is pumped through a circulating system in a large power transformer. The leakage current, the electrical potential at the neutral terminal point of the transformer and the surface electrical potential of the oil are investigated. And the leakage current from the neutral terminal point is measured as a function of a bias polarity applied to a transformer case to investigate the polarty of ion which is absorbed in the case at the interface between the case and oil. As a result, it is found that insulating materials, and it is suggested that the leakage current is the sum of the relaxation current by positive charge from insulating oil to the neutral terminal point and by electrification current from negative charge electrified by the contact with solid insulating materials.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.12C
• /
• pp.1251-1264
• /
• 2002

Currently, audio streaming services such as on-demand service and live service support only a limited number of clients. They also suffer from a lack of stability and degradation of service quality due to their inefficient use of network resources. Futhermore, since the streaming services usually do not consider dynamic services, they are very inconvenience to use. In order to resolve these drawbacks, we propose a novel audio streaming system based on MPEG-2 AAC file format which are facilitated with the network bandwidths efficiently. The proposed system supports QoS for audio streaming as well as guarantees a stability while streaming service is undergoing. Moreover, the system provides a dynamic interface which enables us to use the streaming service more easily and to manage streaming servers with convenient manner. On the contrary, most of the current available static interface streaming services are mainly depending only on a single fixed web page between client and server, which in consequence lead us to use unflexible static service environment. Our implementation results show the proposed system improves the performance compared to those of the currently existing systems that use MP3 file format. It also provides some benefits such as a stability of service and a easy to management of streaming servers.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.10 no.6
• /
• pp.353-362
• /
• 2015

In order to provide smart devices with high quality multimedia streaming services, an adaptive streaming technique over HTTP has been received much attention recently and the Dynamic Adaptive Streaming over HTTP (DASH) standard has been established. In DASH, however, the technique to select an appropriate quality of multimedia based on the performance metrics measured in a smart device might have some difficulties to reflect the capabilities of other neighboring smart devices and dynamic network conditions in real time. To solve the problem, this paper proposes a novel software agent approach, called DASH agent (DA), which gathers and analyzes the device capabilities and dynamic network conditions in real time and finally determines the highest achievable quality of segment to meet the best Quality of Experience (QoE) in current situations. The simulation results show that our approach provides higher quality of multimedia segments with less frequency of quality changes to lower quality of multimedia segments.

• Journal of East Asia Management
• /
• v.1 no.1
• /
• pp.65-87
• /
• 2020

In recent years, with the rapid development of network live streaming, with the popularization of mobile Internet and mobile terminal equipment, the live streaming industry has ushered in great development. A sudden outbreak of the COVID-19 makes the PC end live streaming which has been developed for many years enter a new era, giving birth to the rapid development of mobile end live streaming. Not only because of the expansion of the live streaming industry market, the rise of the trend of the national live streaming, but also because the mobile live streaming is more and more valued by the brand, becoming an important tool for brand communication and product promotion. It is because of its unique communication characteristics that some scholars believe that the era of precision marketing has been opened by live network. Mobile live from the initial fans to reward and promote the brand, to now in the form of live marketing, consumers can "buy while watching". The time period from the understanding of the goods to the final completion of the purchase behavior has been greatly shortened. It is conducive to improving sales volume and brand awareness. Marketing communication through mobile live platform has become a popular way of brand marketing. This paper mainly studies the current situation, methods, problems and development strategies of brand marketing activities with the help of live streaming platform under the background of mobile internet. Taking Tmall live streaming platform as an example, this paper analyzes several ways of brand marketing with the help of live streaming and some universal characteristics of live streaming marketing by using the relevant theories of marketing. In view of the problems existing in live streaming brand marketing, it puts forward relevant Improvement measures. First of all, the paper puts forward the innovation in content and form. Second, the paper suggests that we should make full use of new technologies such as AR and VR to effectively combine with mobile live broadcasting. Third, the paper explores the integration of multiple channels to create intelligent marketing, and further optimize the live interface of mobile terminals. Finally, the paper emphasizes that the government departments and the platform itself should jointly supervise the mobile network live streaming platform and establish a good live broadcasting environment for mobile terminals. With the help of mobile live streaming, the marketing mode has an important impact on the promotion of brand marketing. How to make better use of this business mode and accurately use mobile live broadcast to promote brand marketing, so that enterprises can create greater profits, is also of profound research significance.

• Journal of Environmental Science International
• /
• v.1 no.2
• /
• pp.137-144
• /
• 1992

Flocculation-spectrophotometry and streaming current detector( SCD ) method were investigated and compared in order to determine the optimum dosages of synthetic cationic polymers of different charge density and molecular mass for the removal of humic acid. The optimum dosage for each of the polymers was determined with the dosage at which the lowest absorbance of humic acid was shown for the formal. and was determined with the dosage required during charge neutralization of humic acid for the latter It was in good agreement between both methods and there is a strong inverse correlation between the optimum dosage and charge density of the polymers, with highly charged polymer giving the lowest optimum dosage, pointing out the importance the charge neutralization. By flocculation-spectrophotometry, it was found that the absorbance of humid acid with the amount of each of the polymers dosed, changes sharply for polymers of high charge density, but changes rather broadly for polymers of low and middle charge density, Both methods showed that a stoichiometric correlation exists between the optimum dosage of each of the cationic polymers and the negatively charged humic acid.

• Journal of Institute of Control, Robotics and Systems
• /
• v.10 no.6
• /
• pp.551-556
• /
• 2004

Coagulation process is one of the most important processes in water treatment procedures for stable and economical operation, and coagulant dosing of this process for most plants is generally determined by the jar test. However, this method does not only take a long time to analyze and get the result but also has difficulties in applying to automatic control. This paper shows the feasibility of applying neural network to control the coagulant dosing automatically in water treatment plant. To be specific, the predicted results of the neural network model is shown to be similar to that of jar test. The input variables for learning the neural network are turbidity, water temperature, pH, and alkalinity. Combining the neural network and SCD(Streaming Current Detector) for feedforward and feedback control of injecting coagulant, a rapid change of the raw water quality can be accommodated.

• Environmental Sciences Bulletin of The Korean Environmental Sciences Society
• /
• v.1 no.2
• /
• pp.137-144
• /
• 1997

Flocculation-spectrophotometry and streaming current detector(SCD) method were investigated and compared in order to determine the optimum dosages of synthetic cationic polymers of different charge density and molecular mass for the removal of humic acid. The optimum dosage for each of the polymers was determined with the dosage at which the lowest absorbance of humic acid was shown for the former and was determined with the dosage required during charge neutralization of humic acid for the latter. It was in good agreement between both methods and there is a strong inverse correlation between the optimum dosage and charge density of the polymers, with highly charged polymer giving the lowest optimum dosage, pointing out the importance the charge neutralization. By flocculation-spectrophotometry, it was found that the absorbance of humic acid with the amount of each of the polymers dosed, changes sharply for polymers of high charge density, but changes rather broadly for polymers of low and middle charge density. Both methods showed that a stoichiometric correlation exists between the optimum dosage of each of the cationic polymers and the negatively charged humic acid.

• Journal of Digital Contents Society
• /
• v.9 no.4
• /
• pp.533-541
• /
• 2008

In distributed mobile systems, the congestion control and disconnection problems are current major issues. This paper proposes a grouping-based fuzzy reference streaming method to improve the performance of systems supporting distributed mobile transactions. The proposed method resolves transaction requests issued by mobile clients using the GS interface. In the paper disconnection problems are resolved efficiently using transaction read and update for improved streaming service. Experimental results show that the proposed method outperforms the other existing methods significantly.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.12B
• /
• pp.735-742
• /
• 2007

We propose an effective packet marking algorithm for video streaming in DiffServ network. Because legacy packet markers such as srTCM(single rate three color marker) cannot distinguish the importance of packet, these markers can cause quality degradation of streaming during the network congestion period. Recently proposed TMS (Two Marker System) [4] shows effectiveness in such scenario that video streaming service is struggling with other types of service traffic. However, if many video streaming services co-exist in DiffServ network and result in competition among themselves, then both legacy packet markers and even TMS cannot prevent drastic streaming quality degradation. To cope with this, we suggest A-TCPM (Adaptive time sliding window Three Color Marker) algorithm. In this algorithm, an A-TCPM module decides the color of a racket based upon the probability which is lead by current channel status and frame importance ratio. Simulation results show that proposed A-TCPM algorithm can enhance streaming service quality especially when overbooked video streaming sessions struggle with themselves.