• Journal of Digital Contents Society
• /
• v.10 no.1
• /
• pp.97-105
• /
• 2009

For effective remote education using multimedia, it is necessary to develop efficient management techniques of video information. This requires real-time processing of video information which should be managed and retrieved in a compressed forms. The main technology of compressing video is currently MPEG-2. This implies that it is very important to manage and retrieve video compressed in MPEG-2, and then to process the video in real-time for the remote education environment using multimedia. This paper is to develop the management system of video information which is one of the most critical requirements in remote education systems for managing and retrieving MPEG-2 video.

• Journal of the Semiconductor & Display Technology
• /
• v.13 no.1
• /
• pp.27-33
• /
• 2014

Video from unmanned aerial vehicle (UAV) is influenced by natural environments due to the light-weight UAV, specifically by winds. Thus UAV's shaking movements make the video shaking. Objective of this paper is making a stabilized video by removing shakiness of video acquired by UAV. Stabilizer estimates camera's motion from calculation of optical flow between two successive frames. Estimated camera's movements have intended movements as well as unintended movements of shaking. Unintended movements are eliminated by smoothing process. Experimental results showed that our proposed method performs almost as good as the other off-line based stabilizer. However estimation of camera's movements, i.e., calculation of optical flow, becomes a bottleneck to the real-time stabilization. To solve this problem, we make parallel stabilizer making average 30 frames per second of stabilized video. Our proposed method can be used for the video acquired by UAV and also for the shaking video from non-professional users. The proposed method can also be used in any other fields which require object tracking, or accurate image analysis/representation.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.118-123
• /
• 2009

In this paper, we propose packet selection and significance based interval allocation algorithm for real-time streaming service. In real-time streaming of inter-frame (and layer) coded video, minimizing packet loss does not imply maximizing QoS. It is true that packet loss adversely affects the QoS but one single packet can have more impact than several other packets. We exploit the fact that the significance of each packet loss is different from the frame type it belongs to and its position within GoP. Using packet dependency and PSNR degradation value imposed on the video from the corresponding packet loss, we find each packet's significance value. Based on the packet significance, the proposed algorithm determines which packets to send and when to send them. The proposed algorithm is tested using publicly available MPEG-4 video traces. Our scheduling algorithm brings significant improvement on user perceivable QoS. We foresee that the proposed algorithm manifests itself in last mile connection of the network where intervals between successive packets from the source and to the destination are well preserved.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.7 no.1
• /
• pp.21-26
• /
• 2012

This paper presents a real-time image transferring and localization system utilizing HSDPA, a commercial wireless network system. A novel image compression algorithm is developed based on MPEG4 to comply with uploading bandwidth of 130 kbps and QVGA image transmission of 30 fps. Aiming at being embedded in moving vehicles, the proposed system has a small size, low power consumption, and robustness to disturbances. We validate the performance of the system by presenting captured images of transferring video and localization data. Our system can be applied to real-time surround monitoring in moving vehicles or real-time ecology observation in remote places.

• 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.

• ETRI Journal
• /
• v.44 no.1
• /
• pp.62-72
• /
• 2022

In this study, we propose a multi-GPU-based 8KVR stitching system that operates in real time on both local and cloud machine environments. The proposed system first obtains multiple 4 K video inputs, decodes them, and generates a stitched 8KVR video stream in real time. The generated 8KVR video stream can be downloaded and rendered omnidirectionally in player apps on smartphones, tablets, and head-mounted displays. To speed up processing, we adopt group-of-pictures-based distributed decoding/encoding and buffering with the NV12 format, along with multi-GPU-based parallel processing. Furthermore, we develop several algorithms such as equirectangular projection-based color correction, real-time CG overlay, and object motion-based seam estimation and correction, to improve the stitching quality. From experiments in both local and cloud machine environments, we confirm the feasibility of the proposed 8KVR stitching system with stitching speed of up to 83.7 fps for six-channel and 62.7 fps for eight-channel inputs. In addition, in an 8KVR live streaming test on the 5G MEC/cloud, the proposed system achieves stable performances with 8 K@30 fps in both indoor and outdoor environments, even during motion.

• ETRI Journal
• /
• v.28 no.4
• /
• pp.537-540
• /
• 2006

When multiple ZigBee wireless personal area networks (WPANs) are in close proximity to each other, contentions and collisions in transmissions will lead to increased packet delays. However, there is no existing study on how delay performance would be affected in a crowded real-life environment where each person walking down a busy street would be wearing a ZigBee WPAN. This letter studies the use of ZigBee WPANs in such a real-life environment for real-time heart beat monitoring. To be pragmatic, we derived a mobility pattern from the analysis of a real-life video trace. Then, we estimated the delay performance from the video trace by combining data collected from ZigBee experiments. The results show that the 300 ms packet delay requirement will not be met for only 11% of the time. When failure occurs, it will last for an average duration of 1.4 s.

• 한국정보통신설비학회:학술대회논문집
• /
• 2007.08a
• /
• pp.350-353
• /
• 2007

This paper presents an real-time implementation of H.264/AVC High 4:4:4 Predictive profile decoder using general-purpose processors by exploiting multi-threading technique and Single Instruction Multiple Data (SIMD) instructions without any quality degradation. We analyze differences between the existing High profile and High 4:4:4 Predictive profile decoder, and show various optimization techniques to decode high fidelity and high definition (HD) video in real-time. Simulation results show that the proposed decoder can play high fidelity HD video at average 40 frames per seconds (fps) for the IBBrBP bistream and about 50 fps for the Intra-only bitstream.

• Korean Journal of Computational Design and Engineering
• /
• v.19 no.3
• /
• pp.294-304
• /
• 2014

This paper presents a real-time rendering algorithm of large-scale geometric data using GLSL (OpenGL shading language). It details the VAO (vertex array object) and VBO(vertex buffer object) to be used for up-loading the large-scale point clouds and polygon meshes to a graphic video memory, and describes the shader program composed by a vertex shader and a fragment shader, which manipulates those large-scale data to be rendered by GPU. In addition, we explain the global rendering procedure that creates and runs the shader program with the VAO and VBO. Finally, a rendering performance will be measured with application examples, from which it will be demonstrated that the proposed algorithm enables a real-time rendering of large amount of geometric data, almost impossible to carry out by previous techniques.

• Proceedings of the KSRS Conference
• /
• 2007.10a
• /
• pp.508-511
• /
• 2007

Car navigation system is one of the most important applications in telematics. A newest trend of car navigation system is using real video captured by camera equipped on the vehicle, because video can overcome the semantic gap between map and real world. In this paper, we suggest a visual car navigation system that visually represents navigation information or route guidance. It can improve drivers' understanding about real world by capturing real-time video and displaying navigation information overlaid on it. Main services of the visual car navigation system are graphical turn guidance and lane change guidance. We suggest the system architecture that implements the services by integrating conventional route finding and guidance, computer vision functions, and augmented reality display functions. What we designed as a core part of the system is visual navigation controller, which controls other modules and dynamically determines visual representation methods of navigation information according to a determination rule based on current location and driving circumstances. We briefly show the implementation of system.