• Journal of Korea Multimedia Society
• /
• v.9 no.9
• /
• pp.1139-1149
• /
• 2006

This paper describes a novel and efficient algorithm, which extracts focused objects from still images with low depth-of-field (DOF). The algorithm unfolds into four modules. In the first module, a HOS map, in which the spatial distribution of the high-frequency components is represented, is obtained from an input low DOF image [1]. The second module finds OOI candidate by using characteristics of the HOS. Since it is possible to contain some holes in the region, the third module detects and fills them. In order to obtain an OOI, the last module gets rid of background pixels in the OOI candidate. The experimental results show that the proposed method is highly useful in various applications, such as image indexing for content-based retrieval from huge amounts of image database, image analysis for digital cameras, and video analysis for virtual reality, immersive video system, photo-realistic video scene generation and video indexing system.

• Journal of the Korea Computer Graphics Society
• /
• v.25 no.4
• /
• pp.9-16
• /
• 2019

In this paper, we propose a hydrodynamic scene separation method for wave propagation from video imagery using autoencoder. In the coastal area, image analysis methods such as particle tracking and optical flow with video imagery are usually applied to measure ocean waves owing to some difficulties of direct wave observation using sensors. However, external factors such as ambient light and weather conditions considerably hamper accurate wave analysis in coastal video imagery. The proposed method extracts hydrodynamic scenes by separating only the wave motions through minimizing the effect of ambient light during wave propagation. We have visually confirmed that the separation of hydrodynamic scenes is reasonably well extracted from the ambient light and backgrounds in the two videos datasets acquired from real beach and wave flume experiments. In addition, the latent representation of the original video imagery obtained through the latent representation learning by the variational autoencoder was dominantly determined by ambient light and backgrounds, while the hydrodynamic scenes of wave propagation independently expressed well regardless of the external factors.

• Journal of the Korean Society of Food Culture
• /
• v.35 no.2
• /
• pp.181-192
• /
• 2020

This study aimed to identify whether or not four service situations varying according to positive and negative combinations of physical evidence and functional service influence store image and purchase behavioral intentions of customers at bulgogi restaurants. The video-scenario technique was used for the study. Data were analyzed with the SPSS (Window 19.0) package using frequency analysis, one-way ANOVA, 2 by 2 factorial ANOVA, exploratory factor analysis, and multiple regression analysis to confirm the hypotheses. The combined effect of functional service and physical evidence influenced store image and purchase intention. In terms of seperate effect of physical evidence and functional service, the effect of employee service on store image was more powerful than that of physical evidence, even though the effect differed depending on the situation. Purchase intention was only influenced by functional service quality from employees under the four different scenarios. Thus, when opening a Korean restaurant, proper management of tangible evidence suitable to service, and the prices expected from local customers should be determined. Additionally, extremely high or low levels of physical evidence management should be avoided.

• Journal of Satellite, Information and Communications
• /
• v.9 no.4
• /
• pp.26-31
• /
• 2014

In this paper, we present performance analysis of video smoke detection based on BPN-Network that is using multi-smoke feature, and Neural Network. Conventional smoke detection method consist of simple or mixed functions using color, temporal, spatial characteristics. However, most of all, they don't consider the early fire conditions. In this paper, we analysis the smoke color and motion characteristics, and revised distinguish the candidate smoke region. Smoke diffusion, transparency and shape features are used for detection stage. Then it apply the BPN-Network (Back-Propagation Neural Network). The simulation results showed 91.31% accuracy and 2.62% of false detection rate.

• International Journal of Internet, Broadcasting and Communication
• /
• v.11 no.4
• /
• pp.76-85
• /
• 2019

We introduce a vision-based object detection method for real-time video surveillance system in low-end edge computing environments. Recently, the accuracy of object detection has been improved due to the performance of approaches based on deep learning algorithm such as Region Convolutional Neural Network(R-CNN) which has two stage for inferencing. On the other hand, one stage detection algorithms such as single-shot detection (SSD) and you only look once (YOLO) have been developed at the expense of some accuracy and can be used for real-time systems. However, high-performance hardware such as General-Purpose computing on Graphics Processing Unit(GPGPU) is required to still achieve excellent object detection performance and speed. To address hardware requirement that is burdensome to low-end edge computing environments, We propose sub-frame analysis method for the object detection. In specific, We divide a whole image frame into smaller ones then inference them on Convolutional Neural Network (CNN) based image detection network, which is much faster than conventional network designed forfull frame image. We reduced its computationalrequirementsignificantly without losing throughput and object detection accuracy with the proposed method.

• The Journal of the Korea Contents Association
• /
• v.20 no.9
• /
• pp.29-42
• /
• 2020

The purpose of this study is to present a qualitative research method for extracting and analyzing the comments written by YouTube video users. To do this, we used YouTube users' feedback to measure the hedonic, social, and utilitarian gratification of use and gratification theory(UGT) through by using analysis and topic modeling. The result of the measurement found that the first reason why users watch the trot singer, Cho Myung-sub's video in the KBS Korean broadcasting channel is to achieve hedonic gratification with high frequency. In word-document network analysis, the degree of centrality was high in words, such as 'cheering', 'thank you', 'fighting', and 'best'. Betweenness centrality is similar to the degree of centrality. Eigenvector centrality also shows that words such as 'love', 'heart', and 'thank you' are the most influential words of users' opinions. The results of the centrality analysis present that the majority of video users show their 'love', 'heart' and 'thank you' for the video. it indicates that the high words in centrality analysis is consistent with the high frequency words of hedonic and social gratification dimension of the UGT. The study has research methodological implication that shed light on the motivations for watching YouTube videos with UGT using text mining techniques that automate qualitative analysis, rather than following a survey-based structural equation model.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.11
• /
• pp.1472-1480
• /
• 2016

When disasters such as earthquake, flood, typhoon, fire and terrorism are occurred a video management system is required which can shoot video on the disaster site, and send it to a server located in the command center that can grasp at a glance the site conditions. In this paper, to know the performance elements information needed to construct the video management system. we presented a method of configuring an ad hoc network based on Wi-Fi direct communication. And we also evaluated the performance through built of environment for the video management system. The evaluated performance test items are 'packet flow tests according to the video compression ratios and the image sizes', 'packet flow tests according to the distance and traffic tests', 'packet flow tests according to whether LOS or not', and 'performance test of scenarios'.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.9 no.10
• /
• pp.1197-1205
• /
• 2014

Recently signal transmission of ultra high-definition(4K-UHD) video system is transferred as uncompressed high speed data. However, this has a limit to compose the system because EMI between separate cables of high speed interface section and skew bring distortion of the video signal and jitter. In this paper we applied V-by-One HS interface technique to transfer uncompressed high speed data. We analyzed HSD(High Speed Differential) transmission line signal integrity. Also we applied RF transmission technique instead of UHD video pattern control interface PCB design. When we measured V-by-One HS video signal of designed 4K-UHD class signal generator, We found that the transmission performance has been signal standard.

• Journal of Engineering Education Research
• /
• v.15 no.4
• /
• pp.66-75
• /
• 2012

The purpose of this study is to analyze the learners' satisfactions of the video conferencing for global engineering education and provide the implications for effective video conferencing. This study analyzed the learners' satisfactions on their participation in the video conferencing. We conducted a questionnaire survey with 132 subjects who studied in the video conferencing universities, and performed interviews with 27 learners. The results include learners wanted to have more practical themes to satisfy their interests; faculty roles among domestic professors and abroad professors should be specified: distance learners should be guided to have diverse interactions among themselves. The implications are explored for the future study and video conferencing practices.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.10B
• /
• pp.1971-1978
• /
• 1999

In this paper, we study the effects of long-range dependence (LRD) in VBR video traffic on queueing system. This paper consists of Part I and II. In Part I, we present a (LRD) video traffic model based on the shifting-level (SL) process. We observe that the ACF of an empirical video trace is accurately captured by the shifting-level process with compound correlation (SLCC): an exponential function in short range and a hyperbolic function in long range. We present an accurate parameter matching algorithm for video traffic. In the Part II, we offer the queueing analysis of SL/D/1/K called ‘quantization reduction method’. Comparing the queueing performances of the DAR(1) model and the SLCC with that of a real video trace, we identify the effects of SRD and LRD in VBR video traffic on queueing performance. Simulation results show that Markoivian models can estimate network performances fairly accurately under a moderate traffic load and buffer condition, whereas LRD may have a significant effect on queueing behavior under a heavy traffic load and large buffer condition.