0
1
2
3
4
5
6
7
0
1
2
3
(a) 8-direction (b) 4-direction
1
2
3
An Efficient Contour Tracking and Coding Algorithm
Based on Vertex Chain Code
Wen Li
School of Information Science and Engineering
Lanzhou University
Lanzhou, China
liwen.lzu@gmail.com
Yan Li, Yide Ma, Member, IEEE
School of Information Science and Engineering
Lanzhou University
Lanzhou, China
{liyan.lzu, yidema}@gmail.com
Abstract—Based on analysis of contours of irregular region, and
according to the characteristic that massive continuous code and
the same specific code combination are usually contained in a
region boundary’s vertex chain code, a new effective contour
tracking algorithm and representation method based on the pixel
vertex matrix is proposed. Moreover, we re-encoding the new
vertex chain code using a Huffman coding strategy and then
select the more compressed result as the output. The results
showed that the new method could effectively reduce code stream
for contours, hence increase the compression ratio of the image.
Keywords- image processing; vertex chain code; contour
tracking; pixel vertex matrix; huffman coding
I. INTRODUCTION
Image compression plays an important part in the field of
digital image processing, and the second-generation coding
technique based on human visual properties, including irregular
segmented region coding, is popular due to its high
compression ratio and high quality of image reconstruction.
Human eye is very sensitive to the edges of the image so that
processing the boundaries of regions in an image separately to
compress the information of edges with low compression ratio
is of great help for digital image understanding and analysis.
The chain code is often used in image processing and pattern
recognition for representing the lines, curves or boundaries of
regions. The main reason for the popularity of chain code is its
memory compactness. The first chain code for representing
digital curves was proposed by Freeman [1],which was an
eight-directional encoding scheme for line drawings. Chain
code can be viewed as a connected sequence of straight-line
segments with specified lengths and directions. The direction
between two adjacent points is defined as a link. A link is
consists of eight possible directions which are numbered from
0 to 7. A four-directional version of the Freeman chain code is
also used sometimes, as shown in Fig.1.
In 1999, Ernesto Bribiesca [2] proposed a chain code for
shape representation called Vertex chain code (VCC). VCC
determines the number of regular cells(pixels), which are in
touch with the boundary contour of the shape. Three elements
(1, 2, and 3) are needed to describe the bounding contour of the
shape composed of pixels (an example of VCC is shown in Fig.
2).
Many applications for chain codes have been reported. For
example, Wang et al. [3] applied chain codes for pavement
crack classification. Rehman [4] proposed an algorithm for
nastaliq character representation using a new scale invariant
optimized chain code. Ranjan and Chhanda [5] presented an
image registration algorithm using object shape’s chain code.
Oldooz and Habibollah [6] introduced an algorithm for length
estimation based on the vertex chain code. An efficient
contour-based algorithm for binary video shape coding was
developed by Wang and Lai [7]. Zhou and Zahir [8] proposed a
context-based relative-directional chain coding strategy, which
was efficient for coding the data in one-pass processing method.
In this paper, we proposed a region contour tracking and
coding algorithm based on the pixel vertex matrix of the image,
which can effectively reduce the bit requirement for
maintaining the codes of region contours so that improve the
performance of compression of the whole image.
Figure 1. Freeman chain code Figure 2. Vertex chain code
II. AN EFFICIENT CONTOUR TRACKING AND
CODING ALGORITHM
The boundaries of regions in an image are formed by the
severe intensity changes between adjacent parts of an image.
For irregular region segmented image coding, we expect that
the region is closed so that we can use closed contour to
present the shape of the region. The contour of a region should
be concise in order to meet the need of the follow-up operation
such as storage and transmission. Meanwhile, it should be
accurate to ensure that it shows the feature of an image
precisely. Therefore the effectiveness and accuracy of the
contour of a region are the basic requirements in boundaries
representation.
As we know, chain code scheme can describe the contours
with high accuracy. But there are a large number of redundant,
for it happens most of the time that the direction from current
pixel to the next pixel is usually the same or nearly the same as
the direction from previous pixel to this one [7]. Statistical
2012 Eighth International Conference on Intelligent Information Hiding and Multimedia Signal Processing
978-0-7695-4712-1/12 $26.00 © 2012 IEEE
DOI 10.1109/IIH-MSP.2012.130
514