• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.705-710
• /
• 2006

인터넷의 성장과 함께 인터넷 상의 비디오 스트리밍은 콘텐츠 전달에 중요한 수단이 되었다. 기존의 비디오 전송 방식은 사용자 터미널에 하나의 압축된 비디오를 스트리밍하는 것이었으나 최근 다중 뷰 비디오 코딩은 다른 시각 각도에 비디오 스트리밍 서비스를 지원하는 새로운 어플리케이션이 되었다. 따라서 단일 비디오는 다른 초점을 갖는 여러 독립된 비디오 스트림으로 구성되며 이는 더 많은 대역폭을 요구한다. 하지만 시간에 따라 네트워크의 대역폭이 변하는 경우, 계획된 비트율에 따라 콘텐츠의 비트량을 조절하여 양질의 콘텐츠 제공을 보장하기 어렵다. 따라서 트랜스코딩은 가변적인 네트워크의 대역폭에 적응하기 위한 필수적인 처리 단계가 되었다. 본 논문에서는 동시에 다중 비디오의 선택 및 트랜스코딩을 함께 고려하는 최적화 방법을 기반으로 한 스트리밍 다중 뷰 비디오의 최적 적응 프래임워크를 제안한다. 최적화된 비디오 스트림의 집합은 네트워크 대역폭, 트랜스코딩 코스트와 같은 제한을 만족하고 최적의 인지적 화질을 갖도록 선택되거나 트랜스코딩될 수 있으며 이는 MCKP로 모델링될 수 있다. 또한 트랜스코딩시 R-D 커브 상에 최적의 점을 얻기 위해 분리계획법이 적용하며 결과를 비교 분석한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.1C
• /
• pp.47-55
• /
• 2003

This paper proposes a composition method of mosaic image from the compressed MPEG-2 video stream, in which the displacement between successive frames according to the camera operation is estimated directly from the information involved in the stream. In the proposed method. the approximated optical flow is constructed from motion vectors of macro blocks, and it is used to determine the parameters of the displacements according to the camera operation associated with pan and tilt. The extracted parameters are used to determine the geometric transform of successive video frames in order to construct a mosaic image. The construction of mosaic uses several blending techniques including the one proposed by Nichols in which an analytic weight is used to determine pixel values. Through the experiment, the blending technique based on analytic weights was superior to the others such as averaging and median-based techniques. It provided more smooth changes in background and made use of instantaneous frame information to construct a mosaic. The mosaic in the paper puts the emphasis on the reduction of computation because it is constructed from the motion vectors included in the compressed video without decoding and recalculating exact optical flows. The constructed mosaic can be used in the retrieval of the compressed video as the representative frame of a shot.

• Journal of KIISE:Software and Applications
• /
• v.35 no.4
• /
• pp.207-218
• /
• 2008

With the growth of the Internet, multimedia streaming becomes an important means to deliver video contents over the Internet and the amount of the streaming multimedia contents is also getting increased. However, it becomes difficult to guarantee the quality of service in real-time over the IP network environment with instantaneously varying bandwidth. In this paper, we propose an optimal adaptation framework for streaming contents over the Internet in the sense that the perceptual quality of the multi-angie content with multiple visual objects is maximized given the constraints such as available bandwidth and transcoding cost. In the multi-angle video service framework, the user can select his/her preferred alternate views among the given multiple video streams captured at different view angles for a same event. This enhanced experience often entails streaming problems in real-time over the network, such as instantaneous bandwidth changes in the Internet. In order to cope with this problem, we assume that multi-angle video contents are encoded at different bitrates and the appropriate video streams are then selected or transcoded for delivery to meet such bandwidth constraints. For the user selective consumption of the various bitstreams in the multi-angle video service, the bitstream in each angle can be encoded in various bitrate, and the user can select a sub-bitrstream in the given bitrstreams or transcode the corresponding content in order to deliver the optimally adapted video contents to the instantaneously changing network condition. Therefore, we define the transcoding cost which means the time taken for transcoding the video stream and formulate a unified optimization framework which maximizes the perceptual quality of the multiple video objects in the given constraints such as the transcoding cost and the network bandwidth. Finally, we present plenty of the experimental results to show the effectiveness of the proposed method.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.4
• /
• pp.377-385
• /
• 2015

This paper proposes a novel forward vehicle tracking algorithm based on the WMIL(Weighted Multiple Instance Learning) equipped with a particle filter. In the proposed algorithm Haar-like features are used to train a vehicle object detector to be tracked and the location of the object are obtained from the recognition result. In order to combine both the WMIL to construct the vehicle detector and the particle filter, the proposed algorithm updates the object location by executing the propagation, observation, estimation, and selection processes involved in particle filter instead of finding the credence map in the search area for every frame. The proposed algorithm inevitably increases the computation time because of the particle filter, but the tracking accuracy was highly improved compared to Ababoost, MIL(Multiple Instance Learning) and MIL-based ones so that the position error was 4.5 pixels in average for the videos of national high-way, express high-way, tunnel and urban paved road scene.

• The KIPS Transactions:PartA
• /
• v.9A no.2
• /
• pp.207-216
• /
• 2002

This paper present a design and implementation of a distributed audio, Video stream service framework based on CORBA for efficient processing and control of audio/video stream. We design software components which support processing, control and transmission of audio/video streams as distributed objects. For optimization of stream transmission performance, we separate the transmission path of control data and media data. Distributed objects are defined by IDL and implemented using JAVA. And device dependent facilities like media capturing, playing and communication channels are implemented using JMF (Java Media Framework) components. We show a connection establishment and control procedure of streams communication. And for evaluation, we implement a test system and experiment a system performance. Our experiments show that test system has somewhat longer connection latency time compared to TCP connection establishment, but has optimized media transmission time compared to CORBA IIOP. Also test system show acceptable service quality of media transmission.

• Journal of KIISE:Software and Applications
• /
• v.28 no.5
• /
• pp.389-401
• /
• 2001

This paper proposes an abstraction mechanism of the low-level digital video features for the automatic retrievals of the explosion scenes from the digital video library. In the proposed abstraction mechanism, the regional dominant colors of the key frame and the motion energy of the shot are defined as the primary abstractions of the shot for the explosion scene retrievals. It is because an explosion shot usually consists of the frames with a yellow-tone pixel and the objects in the shot are moved rapidly. The regional dominant colors of shot are selected by dividing its key frame image into several regions and extracting their regional dominant colors, and the motion energy of the shot is defined as the edge image differences between key frame and its neighboring frame. The edge image of the key frame makes the retrieval of the explosion scene more precisely, because the flames usually veils all other objects in the shot so that the edge image of the key frame comes to be simple enough in the explosion shot. The proposed automatic retrieval algorithm declares an explosion scene if it has a shot with a yellow regional dominant color and its motion energy is several times higher than the average motion energy of the shots in that scene. The edge image of the key frame is also used to filter out the false detection. Upon the extensive exporimental results, we could argue that the recall and precision of the proposed abstraction and detecting algorithm are about 0.8, and also found that they are not sensitive to the thresholds. This abstraction mechanism could be used to summarize the long action videos, and extract a high level semantic information from digital video archive.