• Journal of Electrical Engineering and Technology
• /
• v.7 no.1
• /
• pp.115-123
• /
• 2012

A video analysis system used to detect events in video streams generally has several processes, including object detection, object trajectories analysis, and recognition of the trajectories by comparison with an a priori trained model. However, these processes do not work well in a complex environment that has many occlusions, mirror effects, and/or shadow effects. We propose a new approach to a context-aware video surveillance system to detect predefined contexts in video streams. The proposed system consists of two modules: a feature extractor and a context recognizer. The feature extractor calculates the moving energy that represents the amount of moving objects in a video stream and the stationary energy that represents the amount of still objects in a video stream. We represent situations and events as motion changes and stationary energy in video streams. The context recognizer determines whether predefined contexts are included in video streams using the extracted moving and stationary energies from a feature extractor. To train each context model and recognize predefined contexts in video streams, we propose and use a new ensemble classifier based on the AdaBoost algorithm, DAdaBoost, which is one of the most famous ensemble classifier algorithms. Our proposed approach is expected to be a robust method in more complex environments that have a mirror effect and/or a shadow effect.

• Journal of KIISE:Software and Applications
• /
• v.31 no.10
• /
• pp.1320-1331
• /
• 2004

A video adaptation system translates the source video stream into appropriate video stream while satisfying the network and client constraints and maximizing the video quality as much as possible. This paper proposes a semi-dynamic video adaptation scheme, in which several intermediate video streams and the information for the measuring of video quality are generated statically. The intermediate video streams are generated by reducing the resolution of the video stream by a power of two several times, and they are stored as the intermediate video streams on the video server. The statically generated information for the input video stream consists of the degrees of smoothness for each frame rate and the degree of frame definition for each pixel bit rate. It helps to dynamically generate the target video stream according to the client's QoS at run-time as quickly as possible. Experimental result shows that the proposed adaptation scheme can generate the target video stream about thirty times faster while keeping the quality degradation as less than 2% comparing to the target video stream that is totally dynamically generated, although the extra storages for the intermediate video streams are required.

• Journal of Korea Multimedia Society
• /
• v.14 no.1
• /
• pp.44-53
• /
• 2011

H.264/SVC provides scalable video streams which consist of a base layer and one or more enhancement layers. Thus, it can efficiently adapt encoded streams to individual network conditions by dropping some layers of bit streams. However, on a dynamic environment such as the Internet, random packet losses due to network congestion can cause drastic effect on SVC quality. To avoid network congestion, the rate of video streams should be adjusted by carefully selecting a layer of each stream. In this paper, we propose three layer selection algorithms which can avoid network congestion by using the rate-distortion characteristics of streams. Simulation results show that FS(Far-Sighted) algorithm can maximize the overall PSNR value of streams by efficiently using the characteristics of video streams.

• Annual Conference of KIPS
• /
• 2009.04a
• /
• pp.71-74
• /
• 2009

The data-flow paradigm has been employed in various application areas. It is particularly useful where large data-streams must be processed, for example in video and audio processing, or for scientific visualization. A video-based 3D reconstruction system should process multiple synchronized video streams. The system exhibits many properties that can be targeted using a data-flow approach that is naturally divided into a sequence of processing tasks. In this paper we introduce our concept to apply the data-flow approach to a multi-video 3D reconstruction system.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.10C
• /
• pp.975-982
• /
• 2009

In this paper, we propose a novel mixing algorithm for multiple video streams, which can guarantee low complexity and low end-to-end delay. In the propose algorithm, we change the syntax bitstream in the compressed domain. And we mix the H.264/AVC video streams by coded slice unit. The syntax bitstream for a mixing changes according to the arrangement of videos streams in which it organizes the mixing screen layout. Experimental results show more than 4 video streams could be mixed at high speed. Moreover, it could display in the player which it supports the standard H.264/AVC without the picture quality loss.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.5 no.6
• /
• pp.1104-1109
• /
• 2001

In this paper, we describe a method for real-time transmitting video streams based on RTP. In order to guarantee synchronous video streams, we propose a method to grasp network situations by analyzing end-to-end network traffic. In addition, we present an algorithm for satisfying QoS requirements of real-time multimedia transmission and maintaining continuity of transmission. It describes a buffering method for overcoming bandwidth limitations and an analyzing method based on RTCP for grasping network traffic situation to resolve the problem of real-time transmission of video streams.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2001.10a
• /
• pp.169-172
• /
• 2001

In this paper, we describe a method for real-time transmitting video streams based on RTP In order to guarantee synchronous video streams, we propose a method to grasp network situations by analyzing end-to-end network traffic. In addition, we present an algorithm for satisfying QoS requirements of real-time multimedia transmission and maintaining continuity of transmission. It describes a buffering method for overcoming bandwidth limitations and an analyzing method based on RTCP for grasping network traffic situation to resolve the problem of real-time transmission of video streams.

• Journal of Information Processing Systems
• /
• v.19 no.4
• /
• pp.439-449
• /
• 2023

This paper deals with a resource efficient artificial intelligence (AI) service architecture for multi-channel video streams. As an AI service, we consider the object detection model, which is the most representative for video applications. Since most object detection models are basically designed for a single channel video stream, the utilization of the additional resource for multi-channel video stream processing is inevitable. Therefore, we propose a resource efficient AI service framework, which can be associated with various AI service models. Our framework is designed based on the modular architecture, which consists of adaptive frame control (AFC) Manager, multiplexer (MUX), adaptive channel selector (ACS), and YOLO interface units. In order to run only a single YOLO process without regard to the number of channels, we propose a novel approach efficiently dealing with multi-channel input streams. Through the experiment, it is shown that the framework is capable of performing object detection service with minimum resource utilization even in the circumstance of multi-channel streams. In addition, each service can be guaranteed within a deadline.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.1
• /
• pp.95-101
• /
• 2005

In this paper, some performance characteristics of multimedia traffic for mobile ad-hoc networks is studied with simulations. Synthetic streaming video is considered as the multimedia traffic for MPEG-coded video in the simulation. The synthetic video stream is generated with a video stream generation algorithm. The algorithm generate VBR traffics for MPEG video streams with special predefined GOP(group of pictures) patterns that is consisted of a sequence of I(intra-coded), P(predicted-coded) and B(bidirectional-coded) frames. The synthetic VBR streams is transmitted through UDP protocol with on-demand mobile ad-hoc network routing protocols like as AODV and DSR. And performances for video streams through mobile ad-hoc networks is evaluated, the throughputs is compared between data and video traffics.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.8
• /
• pp.3121-3135
• /
• 2015

Owing to the increase in consumption of multimedia content and the improvement of device capacity, user demand for high-quality content has increased. However, it is difficult to transport such large amounts of content over the existing broadcasting network with limited bandwidth. To provide high-definition broadcasting, some studies suggest methods of transporting multimedia over heterogeneous networks after encoding content hierarchically. MPEG Media Transport (MMT), standardized by Moving Picture Experts Group (MPEG), is a solution that enables large-volume media transport over heterogeneous networks such as digital broadcasting networks and packet-switched networks. In the case of delivering a scalable encoded video over different networks, synchronization of each stream is an important issue. MMT defines a synchronization scheme, but does not contain sufficient functions to implement it. In this paper, we propose a synchronization scheme for media streams that are encoded hierarchically, divided into layers, and transported over heterogeneous networks. We implement our scheme using MMT and HTTP, and experimented using three encapsulated video streams with different durations. As a result, we show that the proposed scheme can reduce the waiting time to display high-quality video, relative to Dynamic Adaptive Streaming over HTTP-Scalable Video Coding (DASH-SVC) by requesting segments of enhanced layers after calculating the transmission time. Additionally, we find out that the selection of durations have a relation to the characteristics of the video.