06/30/25 N.
NANDHAGOPAL,L/ECE
IMAGE REPRESENTATION
In image analysis the first step is image
segmentation then the next step is image representation
06/30/25
and description methods.
The output of the segmentation process is the
N.NANDHAGOPAL,L/ECE
boundary pixels of objects. The representation of object
boundaries is very important.
Generally two methods are used for image
representation
External characteristics of the object.
Internal characteristics of the object.
The description is the shape of an image is
needed to recognize one object from another. The
06/30/25
discrimination of a rectangular object from a
circular object can be easily done by using the shape
N.NANDHAGOPAL,L/ECE
of the object contour.
An object with a smooth reflect surface can be
easily separated from an object with a rough surface
by comparing their surface texture so region and
contour methods of describing objects within an
image are very important. The edge of an object
gives its size and shape.
06/30/25 N.NANDHAGOPAL,L/ECE
CHAIN CODE
The method of chain code was introduced in
1961 by freeman.
06/30/25
Chain code describe an object by a sequence of
unit size.
N.NANDHAGOPAL,L/ECE
“Chain codes are used to represent a boundary
by a connected sequence of straight line segments of
specified length and direction. ”
This representation is based on four or eight
connectivity of the segments.
In this approach an arbitrary curve is
represented by a sequence of small vector of unit
length and limited set of possible directions.
06/30/25
It describes an object by a sequence of unit size
N.NANDHAGOPAL,L/ECE
line segments with a given orientation.
Connectivity defines the similarity between the
gray level of neighboring pixels, there are three types
of connected pixels which describe how two
neighboring pixels are related, they are
4- Directional chain code.
8- Directional chain code.
4- Directional chain code.
06/30/25
N.NANDHAGOPAL,L/ECE
8- Directional chain code.
06/30/25
Digital image usually acquired and processed
in a grid form with spacing in the x and y direction.
N.NANDHAGOPAL,L/ECE
So a chain code can be generated by following
a boundary in say a clockwise direction and assigning
direction with number method.
An approach frequently used to resample the
boundary by selecting a large grid spacing,
06/30/25
N.NANDHAGOPAL,L/ECE
then the boundary is traversed a boundary point
is assigned to each node of the large grid.
06/30/25
N.NANDHAGOPAL,L/ECE
The coarser boundary point represented by an
4-code chain code of a boundary depends on the
06/30/25
straight point.
N.NANDHAGOPAL,L/ECE
The coarser boundary point represented by an
8-code the chain code of a boundary depends on the
06/30/25
straight point.
N.NANDHAGOPAL,L/ECE
06/30/25
The first difference of the chain code instead of
N.NANDHAGOPAL,L/ECE
the code itself. This obtained by counting the number of
direction changes.
The accuracy of the resulting code depends on
the grid spacing,
if the grids are closed then the accuracy is good.
DISADVANTAGES:
06/30/25
The chain code will be varied depends on the
N.NANDHAGOPAL,L/ECE
starting point. This method has two problems,
The resulting chain of codes is long.
If there is any small distances along the
boundary due to noise.
