Page 116 - Video Coding for Mobile Communications Efficiency, Complexity, and Resilience

Chapter 4

Basic Motion Estimation Techniques

4.1 Overview

Motion estimation is an important process in a wide range of disciplines and
applications, such as image sequence analysis, computer vision, target track-
ing, and video coding. Di erent disciplines and applications have di erent
requirements and may, therefore, use di erent motion estimation techniques.
This chapter reviews some basic motion estimation techniques developed
speci cally for video coding. It then carries out a comparative study between
the di erent techniques. The chapter also presents the results of an investiga-
tion into the e!ciency of block-matching motion estimation at very low bit
rates. In particular, the investigation shows that the added complexity of this
process is justi able at such bit rates.
Section 4.2 gives a brief introduction to the basics of motion estima-
tion. Sections 4.3– 4.6 brie*y review the di erential, pel-recursive, frequency-
domain, and block-matching motion estimation methods. Section 4.7 presents
the results of a comparative study of the reviewed techniques, whereas
Section 4.8 investigates the e!ciency of motion estimation at very low bit
rates. The chapter concludes with a discussion in Section 4.9.

4.2 Motion Estimation

As already discussed in Chapter 2 (Section 2.7.2), the most commonly used
video coding method is motion-compensated coding. In the rst stage of this
method, called motion estimation (ME), the motion of objects between a
reference frame and the current frame is estimated. This motion information
is then used in the second stage, called motion compensation (MC), to move
the objects of the reference frame to provide a prediction for the current

93