• IEIE Transactions on Smart Processing and Computing
• /
• v.4 no.2
• /
• pp.110-114
• /
• 2015

This paper reports a novel object-detection technique in video sequences. The proposed algorithm consists of detection of objects in illuminated and occluded videos by using object features and a neural network technique. It consists of two functional modules: region-based object feature extraction and continuous detection of objects in video sequences with region features. This scheme is proposed as an enhancement of the Lowe's scale-invariant feature transform (SIFT) object detection method. This technique solved the high computation time problem of feature generation in the SIFT method. The improvement is achieved by region-based feature classification in the objects to be detected; optimal neural network-based feature reduction is presented in order to reduce the object region feature dataset with winner pixel estimation between the video frames of the video sequence. Simulation results show that the proposed scheme achieves better overall performance than other object detection techniques, and region-based feature detection is faster in comparison to other recent techniques.

• Proceedings of the IEEK Conference
• /
• 2000.07a
• /
• pp.322-325
• /
• 2000

In a batch scheduling policy being different from real video system, the requests are not served immediately due to grouping user's requests upon every scheduling points. Such waiting delays by inefficient managements makes an unfair service to users and increases the possibility of higher reneging rates. This paper proposes an adaptive batch scheduling scheme which reduces the average waiting time of user’s requests and reduces the starvation problem for requesters of less popular movies. The proposed scheme selects dynamically multiple videos in given intervals based on the service patterns which reflect the popularity distribution and resource utilizations. Experimental simulation shows that proposed scheme improves about 20-30 percent of average waiting time and reduces significantly the starving requesters comparing with those of conventional methods such as FCFS and MQL.

• Journal of information and communication convergence engineering
• /
• v.9 no.1
• /
• pp.110-116
• /
• 2011

The production environment of fast and accurate news video has been continuously changing together with technology advance. Especially, the editing of the news video advanced focusing on the non-linear system by the introduction of digital technology and the recording of news video advanced to 'tapeless' base. In this new news video production environment, the recorded news videos are transmitted, edited and archived through the NPS (Network Production System), which dramatically improves the efficiency of the news video production. However, there are some issues which need to be resolved in the introduction of NPS. The purpose of this research is to identify the necessity of NPS introduction in the changing news video production environment, explore the issues which should be resolved and suggest the proper way to introduce the NPS to broadcasting companies.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.05a
• /
• pp.107-108
• /
• 2016

Our study proposes the methods of distinguishing vehicle types using the interval and size of the car. The car videos converts the basic RGB model to Gray model for use and through Canny Edge Direction, it eliminates the background of the car while obtaining feature points through the detection of contours.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.364-368
• /
• 2018

In this paper, we use Faster R-CNN that is one of object detection algorithm and OpenCV that purposes computer vision, to implement the system that can detect danger when a vehicle attempts to change lanes into its own lane in videos of highway, national road, general road and etc. Also, the performance of implemented system is evaluated to prove that the performance is not bad.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2007.08a
• /
• pp.814-817
• /
• 2007

We present a hybrid hardware architecture capable of encoding and decoding a full HD resolution video with 60 fps. A number of technical modifications are applied to an existing image compression algorithm not only to accommodate large videos at a high frame rate but to enhance the quality of synthetic images, such as characters and video game images. Image quality of the proposed algorithm at a 1/6 compression ratio was comparable to that of the BTC based 1/3 compression algorithm.

• Journal of Information Technology Applications and Management
• /
• v.27 no.5
• /
• pp.23-33
• /
• 2020

As the 4G and 5G Internet technologies become more common and developed, an environment for uploading and watching videos is created and spread, in addition to simply uploading posts. Watching and sharing daily life among media contents called Vlog, are very common more than ever. This means that individual users could access Vlog easily and the situation could be new trend. Additionally, academic research about Volg is increasing. We analyzed three independent variables affecting a perceived enjoyment we set up the dependent variable. Information search, self-expression, and social need are set as independent variables and social interaction is set as the control variable. Information search and self-expression are significant effect to perceived enjoyment except social need. In particular, social interaction as a control variable has effect to all relationships.

• Journal of the Semiconductor & Display Technology
• /
• v.9 no.4
• /
• pp.99-102
• /
• 2010

In this study, the 3- dimensional (3-D) learning contents are suggested using 'Augmented Reality' instead of existing 2-dimensional (2-D) learning methods. At the present, there are some 2-D learning methods using texts, image, pictures, and videos called e-learning. However, these one-way 2-D methods have some disadvantages such as declining learner's immersion and concentration. Thus, the 3-D learning contents using 'Augmented Reality' are suggested to compensate the disadvantages. According to the development of information technology (IT), the augmented reality has many applications to the era of ubiquitous. However, there are some disadvantages when learners use these contents as following; non-augmenting by partially hiding from makers and declining concentration by patterns of the makers. In this study, the beneficial marker which can solve this non-augmenting phenomenon is suggested.

• Journal of Institute of Control, Robotics and Systems
• /
• v.21 no.11
• /
• pp.996-1002
• /
• 2015

In this paper, we present an effective system for the 3D scene labeling of objects from RGB-D videos. Our system uses a Markov Random Field (MRF) over a voxel representation of the 3D scene. In order to estimate the correct label of each voxel, the probabilistic graphical model integrates both scores from sliding window-based object detectors and also from object location prior maps. Both the object detectors and the location prior maps are pre-trained from manually labeled RGB-D images. Additionally, the model integrates the scores from considering the geometric constraints between adjacent voxels in the label estimation. We show excellent experimental results for the RGB-D Scenes Dataset built by the University of Washington, in which each indoor scene contains tabletop objects.

• Proceedings of the IEEK Conference
• /
• 2007.07a
• /
• pp.255-256
• /
• 2007

We propose a frame loss concealment algorithm for multi-view video sequences based on correspondence matching, which can hide the effects of frame losses efficiently. To achieve high PSNR performances, we employ a block error concealment scheme to refine the concealed results. Simulation results demonstrate that the proposed algorithm effectively protects the quality of reconstructed videos against transmission errors.