Computationally Efficient MCTF for MC-EZBC Scalable Video Coding Framework
The discrete wavelet transforms (DWTs) applied temporally under motion compensation (i.e. Motion Compensation Temporal Filtering (MCTF)) has recently become a very powerful tool in scalable video compression, especially when implemented through lifting. T
- PDF / 478,815 Bytes
- 8 Pages / 430 x 660 pts Page_size
- 110 Downloads / 224 Views
Abstract. The discrete wavelet transforms (DWTs) applied temporally under motion compensation (i.e. Motion Compensation Temporal Filtering (MCTF)) has recently become a very powerful tool in scalable video compression, especially when implemented through lifting. The major bottleneck for speed of the encoder is the computational complexity of the bidirectional motion estimation in MCTF. This paper proposes a novel predictive technique to reduce the computational complexity of MCTF. In the proposed technique the temporal filtering is done without motion compensation. The resultant high frequency frames are used to predict the blocks under motion. Motion estimation is carried out only for the predicted blocks under motion. This significantly reduces the number of blocks that undergoes motion estimation and hence the computationally complexity of MCTF is reduced by 44% to 92% over variety of standard test sequences without compromising the quality of the decoded video. The proposed algorithm is implemented in MC-EZBC, a 3D-subband scalable video coding system. Keywords: Motion Estimation, Motion Compensated Temporal Filtering, Temporal Filtering, MC-EZBC.
1
Introduction
The Scalable Video Coding (SVC) is one of the most important features of modern video communication system. For a truly scalable coding, the encoder needs to operate independently from the decoder, while in predictive schemes the encoder has to keep track and use certain information from the decoder’s side (typically, target bit-rate), in order to operate properly. The 3D sub-band video coding has appeared recently as a promising alternative to hybrid DPCM video coding techniques; it provides high energy compaction, scalable bit-stream for network and user adaptation and resilience to transmission errors. While early attempts to apply separable 3D wavelet transform directly to the video data didn’t produce high coding gains, it was soon realized that, in order to fully exploit inter-frame redundancy, the temporal part of the transform must compensate for motion between frames. In one of the first attempts to incorporate motion into 3D wavelet video coding, Taubman and A. Ghosh, R.K. De, and S.K. Pal (Eds.): PReMI 2007, LNCS 4815, pp. 666–673, 2007. c Springer-Verlag Berlin Heidelberg 2007
Computationally Efficient MCTF for MC-EZBC SVC Framework
667
Zakhor [1] pre-distorted the input video sequence by translating frames relative to one another before the wavelet transform so as to compensate for camera pan. Wang et. al. [2] used mosaicing to warp each video frame into a common coordinate system and applied a shape-adaptive 3D wavelet transform on the warped video. Both of these schemes adopt a global motion model that is inadequate for enhancing the temporal correlation among video frames in many sequences with local motion. To overcome this limitation, Ohm [3] proposed local block-based motion, similar to that used in standard video coders, while paying special attention to covered/uncovered and ”connected/unconnected” regions. Failing to achieve per
Data Loading...
