A Simple and Efficient Search Algorithm For Block-Matching

This document describes a simple and efficient search algorithm called the Simple and Efficient Search (SES) algorithm for block matching motion estimation. The SES algorithm improves on the Three Step Search (TSS) algorithm by reducing the number of checking points during block matching while maintaining computational complexity and performance comparable to TSS. Specifically, the SES algorithm first checks points in orthogonal directions, then restricts the search area to a quadrant containing the minimum point, and determines step size using equations to further reduce checking points.

Uploaded by

SarathKrishnan

Available Formats

Download as PPTX, PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

18 views

A Simple and Efficient Search Algorithm For Block-Matching

Uploaded by

SarathKrishnan

Available Formats

Download as PPTX, PDF, TXT or read online on Scribd

You are on page 1/ 11

A Simple and Efficient Search Algorithm

for Block-Matching Motion Estimation

Midhila M.
Three step search method
simplicity
significant computational reduction
good performance
has been widely used in real-time video
applications

The principle which supports the TSS is based

on the unimodal error surface assumption.
Simple and efficient search algorithm

To improve the performance while keeping the

computational complexity similar to the original
TSS.

speed up the TSS by a factor of two and maintain

its regularity and good performance comparable
with the TSS.
Block-matching motion estimation is performed
between the current frame and a previously
processed frame of a video sequence.

The current frame is divided into non

overlapped square blocks of pixels with size N*N
and search area is (2W + N + 1)* (2W + N + 1)

( w is the maximum displacement in the pixels

along both horizontal and vertical directions)
look for the block in the search area that
best matches the current block.

The matching criterion is based on the

mean absolute difference (MAD), which
can be expressed as:

Fc - pixel intensity in the current frame

Fp pixel intensity in the previously processed frame
(k, l) is the coordinates of the upper left corner
pixel of the current block

(x; y) represents the displacement in pixels

which is relative to the position of current block

After checking each location in the search area,

the mv is then determined as the (x; y) at which
the MAD has the minimum value, i.e., minimum
error.
TSS, there are a total of eight search
directions.
each direction has an opposite direction which
is also searched.
Minimum value cant be identified from them.
In SES method first check two arbitrary points in
orthogonal directions

Then restrict the search area of each step in

certain quadrant which contains only three
search directions.

The SES algorithm uses the following equations

to determine the number of steps required and
the step-size used.
L = [log2(W + 1)]
ss(n)=2^(L-n)
The first phase is to select a search quadrant
The second phase is to find the minimum error
location in the selected quadrant.

In the first phase, we compute the MADs of

A, B, and C
Based on the above value the quadrant can be
determined.
The total number of checking points for each
block-matching is further reduced compared with
the TSS.
It can be shown that the proposed SES can
further speed up the TSS by a factor about two.