Main Menu
About us
Project Description
Quantitative Results
Research Lines
Research Results
Impact on Society
Press room
Contact us
Secure Login
Events Calendar
« < April 2019 > »
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 1 2 3 4 5

Activity 7: Heterogeneous Video Transcoding

Leader: Pedro Cuenca; Researchers: Jose Luís Martínez, Gerardo Fernández, Rafael Rodríguez,  Antonio Garrido, Francisco Quiles

For Activity 7, It has been changed its name and contents.

1. Brief Description of the Goals

At present, the number of multimedia applications making use of video transcoding systems is growing at an exponential rate. From the point of view of the user, the underlying architecture in this type of applications should be totally transparent.  The main user concern is the service to be provided by such systems to him, in particular the video quality and the error-free and timely transmission of the video material.

While it is expected that in a near future the H.264 video standard will replace the use of MPEG-2 systems, there is still a long way to develop efficient techniques to make possible video transcoding between MPEG-2 and H.264. The research in the field of video transcoding, is particularly interesting due to the ample diffusion and use of MPEG-2 at the present time, and a clear interest by the industry on technologies to facilitate the migration possible from MPEG-2 to H.264. The coexistence of both technologies until the complete adoption of the new standard creates the necessity of technologies of video transcoding, for the conversion of the sequences from MPEG-2 to H.264.

On the other hand, low cost video encoders enable the use of video encoding in low cost devices such as toys, low-power surveillance, wireless cameras, mobile phones and disposable video cameras; all these devices can be formed by a low cost video camera which can be supported with reduced complexity encoding algorithms. However, it is important to emphasize that there is still a notable technological gap between an ideal platform and the available solutions, mainly in the form of the encoder-decoder complexity balance and the associated design costs.

This research line focuses on devising new video coding techniques. The work has focused in two main activities: new proposals for heterogeneous video transcoding in real-time and design aspects of Wyner-Ziv coding, making significant contributions towards building a practical video codec with a view to commercial developments and standardization.

2. Scientific and Technical Developed Activities

The first scientific activity has been to create a Low Complexity MPEG-2 to H.264 Transcoder. The transcoder can reduce extremely computational cost up to 95 % on average with a negligible PSNR degradation (less than 0.3 dB) and slight increase in bit rate (8% on average). The proposed transcoder is for heterogeneous transcoding from MPEG-2 to H.264 in the spatial domain. For fast transcoding, three approaches, that reuse the information incoming for the MPEG-2 decoding stage (DC coefficients, MB mode for a macroblock, the code block pattern and motion vectors), and the MC information residual processed as the mean and the variance, were presented. The macroblock mode is adaptively selected using the Intra and Inter approaches, and also reuse the motion vector to reduce the search range window size. These reductions of the transcoder complexity help in real-time implementation of a Low Complexity MPEG-2 to H.264 Transcoder. This scientific activity has been registered as US patented "Video encoding with reduced complexity" jointly with Prof. Hari Kalva, Multimedia Group, Florida Atlantic University (US).    

In order to efficiently support mobile-to-mobile as well as low cost video communications between low complexity devices, one of the approaches of this activity has been to develop a transcoder framework where traditional video coding such as H.264 and the recently proposed Wyner-Ziv video coding paradigm can operate together. These transcoder devices convert the bit stream between both paradigms in order to leave the low complexity algorithm to end user devices.

For existing contents already encoded in H.264/AVC without any scalability can benefit from scalability is necessary an efficient H.264/AVC-to-SVC transcoder to perform this transcoding faster than if the video is decoded and fully re-encoded again in SVC. This is possible by reusing the information collected from the H.264/AVC decoding stage for acceleration parts of the SVC encoding. Several techniques for accelerating those parts in an H.264/AVC-to-SVC transcoder with temporal scalability have been also proposed in this activity. The main contributions on these activities have derived in a contract enterprise with OVERON S.L. (Spain) and AICOX Solution S.A. (Spain).

Publications: [Gfe07] [Gfe08a] [Gfe08b] [Gfe08c] [Jlm09a] [Jlm09b

Projects funded by Public Calls:  PEII09-0037-2328 and PAI06-0106-7725 by regional grants.

External collaborations Academia:  Hari Kalva, W.A.C. Fernando, Rik Van de Walle 

External collaborations Industry: --

Company Agreements: OVERON S.L. (Spain) and AICOX Solution S.A. (Spain)

PhD dissertations: Jose Luís Martínez MartínezGerardo Fernández Escribano 

Patents:  Video Encoding with Reduced Complexity, 2008