Ziyuan Pan

UMSM: A Traffic Reduction Method on Multi-View Video Streaming for Multiple Users

Multi-view video consists of multiple video sequences captured simultaneously by multiple closely spaced cameras and enables users to freely change their viewpoints by playing different video sequences. Therefore, the transmission bitrate of multi-view video requires more bandwidth than conventional multimedia. In order to reduce the bandwidth requirement, User Dependent Multi-view Video Transmission (UDMVT), which is based on Multi-view Video Coding (MVC), has been proposed for single users. In UDMVT, the same frames are encoded into different versions for each user, which increases the transmission bitrate for multiple users because of the redundant transmission of overlapping frames. In order to address this problem, we proposed User dependent Multi-view video Streaming for Multi-users (UMSM). UMSM possesses two characteristics: overlapping frames required by multiple users are transmitted only once by multicast while un-overlapping frames required by each user are transmitted by unicast and offset of video requests between multiple users is aligned so as to maximize the area of overlapping frames. UMSM achieves a low transmission bitrate of multi-view video for multiple users by means of combining these characteristics. To investigate the effect of users view switching on traffic reduction, we design three types of view-switching models, namely, scanning model, random model, and watching model. Simulation results obtained using benchmark test sequences provided by MERL reveal that UMSM decreases the transmission bitrate by 80.6% on average for five users, as compared to MVC, when users gaze a viewpoint for a while after switching views in order to search an attractive viewpoint.

User Dependent Scheme for Multi-View Video Transmission

The traffic of multi-view video is several times larger than traditional multimedia, which brings much more increment in the bandwidth requirement. To overcome this problem, Simulcast and MVC (Multi-view Video Coding) can be used to decrease the traffic for multi-view video. However, although Simulcast removes redundant information within each view, it still contains a large amount of inter-view redundant information. MVC further removes the correlation between the views, but all the views should be encoded and transmitted. So the bit-rates are still too high in these schemes. In this paper, in order to reduce the bit-rate for multi-view video transmission, we analyze the users motion which is classified into random access model and successive motion model. We find that in the successive motion model only part of frames are possible to be displayed in a period of time. According to the analytical result, we propose a user dependent scheme to decrease the bit-rate of multi-view video transmission. The proposed scheme only encodes and transmits those frames that are possible to be displayed according to the users motion. Further, in order to support the proposed scheme, we also propose a prediction structure as a substitute of the prediction structure in MVC. Performance evaluation proves that this proposed scheme reduces the average bit-rate for the transmission of multi-view video.

A User Dependent System for Multi-view Video Transmission

Multi-view video consists of multiple views which are taken by multiple cameras from different positions and angles. It allows the users to freely change their viewpoints. Typical applications of multi-view video include Free Viewpoint TV, remote medical surgery, wireless multimedia sensor networks and so on. However, the traffic of multi-view video is also several times larger than traditional multimedia, which brings much more increment in the bandwidth requirement. To overcome this problem, Simulcast and MVC (Multi-view Video Coding) can be used to decrease the traffic for multi-view video. However, although Simulcast removes redundant information within each view, it still contains a large amount of inter-view redundant information. MVC further removes the correlation between the views, but all the views should be encoded and transmitted. So the bit-rates are still too high in these schemes. In this paper, in order to reduce the bit-rate for multi-view video transmission, we analyze the users motion which is classified into random access model and successive motion model. We find that in the successive motion model only part of frames are possible to be displayed in a period of time. According to the analytical result, we propose a user dependent system to decrease the bit-rate of multi-view video transmission. The proposed system only encodes and transmits those frames that are possible to be displayed according to the users motion. Further, in order to support the proposed system, we also propose a prediction structure as a substitute of the prediction structure in MVC. Performance evaluation proves that the proposed system reduces the average bit-rate for the transmission of multi-view video and is to be more effective when the number of views is large.

Traffic Reduction for Multiple Users in Multi-view Video Streaming

Multi-view video consists of multiple video sequences captured simultaneously from different angles by multiple closely spaced cameras. It enables the users to freely change their viewpoints by playing different video sequences. Transmission of multi-view video requires more bandwidth than conventional multimedia. To reduce the bandwidth, UDMVT (User Dependent Multi-view Video Transmission) based on MVC (Multi-view Video Coding) has been proposed for single user. In UDMVT, for multiple users the same frames are encoded into different versions for each user, which increases the redundant transmission. For this problem, this paper proposes UMSM (User dependent Multi-view video Streaming for Multi-users). UMSM possesses two characteristics. The first characteristic is that the overlapped frames that are required by multiple users are transmitted only once using the multicast to avoid unnecessary duplication of transmission. The second characteristic is that a time lag of the video request by multiple users is adjusted to coincide with the next request. Simulation results using benchmark test sequences provided by MERL show that UMSM decreases the transmission bit-rate 55.3% on average for 5 users watching the same multi-view video as compared with UDMVT.

Tradeoffs Among Delay, Energy and Accuracy of Data Aggregation for Multi-View Multi-Robot Sensor Networks

© 2012 Li et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Tradeoffs Among Delay, Energy and Accuracy of Data Aggregation for Multi-View Multi-Robot Sensor Networks

Bandwidth-efficient user dependent transmission for multi-view video

Multi-view video allows user to watch 3D video and select desired viewpoint. In order to reduce the switching delay, many researches tend to transmit all the views, which brings much more increment in the bandwidth requirement. And the traffic increase as the increase of the number of the views in the multi-view video. In order to overcome these problems, we have proposed a user dependent scheme called UDMVT for the transmission of multi-view video in [1]. And we further improved this scheme to support the 3D multi-view video in the [2]. The evaluation results proved that UDMVT is efficient to reduce the bit-rate of the transmission of the multi-view video in the successive motion model especially when the number of view is larger. In this paper, we will simply introduce the UDMVT and discuss the application of the UDMVT in the Free Viewpoint TV.

A user-driven system for multi-view video in successive motion model

Recently, multi-view video that is taken by multiple cameras from different positions and angles to provide the real world experience has attracted more attentions. Its typical applications include Free Viewpoint TV, remote medical surgery, wireless multimedia sensor networks and so on. The size of multi-view video is several times bigger than traditional multimedia, which brings much more increment in the bandwidth requirement. Compression technologies, such as MPEG and MVC, can greatly decrease the size of multi-view video. However, as just one view is displayed at a specific time, even with MVC bit-rate of multi-view video is still high. In this paper, we researched the switch models of the multi-view video applications and proposed a user-driven multi-view video streaming system which can significantly decrease the bandwidth requirement and provide better user experience. In the proposed system, only those frames that are possible to be displayed are encoded and transmitted and the data are transported by different protocols to decrease the effect of the network congestion. In order to support this solution, we also improved the prediction structure as a substitute of the prediction structure in MVC. Evaluation proves that this proposed solution is great helpful to reduce the average bit-rate and the bandwidth requirement for the transmission of multi-view video.