Scheduled System Maintenance:
On May 6th, system maintenance will take place from 8:00 AM - 12:00 PM ET (12:00 - 16:00 UTC). During this time, there may be intermittent impact on performance. We apologize for the inconvenience.
By Topic

On the Information Theoretic Performance Comparison of Causal Video Coding and Predictive Video Coding

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

The purchase and pricing options are temporarily unavailable. Please try again later.
3 Author(s)
En-hui Yang ; Dept. of Electr. & Comput. Eng., Univ. of Waterloo, Waterloo, ON, Canada ; Lin Zheng ; Da-ke He

Causal video coding is a coding paradigm where video source frames X1, X2,..., XN are encoded in a frame-by-frame manner, the encoder for each frame can use all previous source frames and all previous encoded frames, and the corresponding decoder can use only all previous encoded frames. In the special case where the encoder for each frame Xk is further restricted to enlist help only from all previous encoded frames, causal video coding is reduced to predictive video coding, which all MPEG-series and H-series video coding standards proposed so far are based upon. In this paper, we compare the rate distortion performance of causal video coding with that of predictive video coding from an information theoretic perspective by modeling each frame Xk itself as a source Xk={Xk(i)}i=1. Let Rc*(D1,...,DN) (Rp*(D1,...,DN), respectively) denote the minimum total rate required to achieve a given distortion level D1,...,DN in causal video coding (predictive video coding, respectively). We first show that like Rc*(D1,..., DN), for jointly stationary and totally ergodic sources X1, X2,..., XN, Rp*(D1,...,DN) is equal to the infimum of the nth order total rate distortion function Rp,n(D1,...,DN) over all n, where Rp,n(D1,...,DN) itself is given by the minimum of an information quantity over a set of auxiliary random variables. We then prove that if the jointly stationary and totally ergodic sources X1,..., XN form a (first-order) Markov chain, we have Rp*(D1,...,DN)=Rc*(D1,...,DN). However, this is not true in general if X1,..., XN do not form a (first-order) Markov chain. Specifica- ly, we demonstrate that for independent and identically distributed vector source (X1,..., XN), if X1,..., XN do not form a (first-order) Markov chain, then under some conditions on source frames and distortion, Rc*(D1,..., DN) is strictly less than Rp*(D1,..., DN) in general. Our techniques allow us to compare Rp*(D1,..., DN) with Rc*(D1,..., DN) even when the single-letter characterization of Rp*(D1,..., DN), if any, is unknown.

Published in:

Information Theory, IEEE Transactions on  (Volume:60 ,  Issue: 3 )