A DETAIL PRESERVED COMPRESSION BASED ON CONTOURLET TRANSFORM'
FeiYan Zhang, LeiGuang Wang, Feng Yuan, Qianqing Qin
State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, 129
Luoyu Road, Wuhan, China, 430079 -
fifi012@163.com
KEY WORDS: Wavelet Based Contourlet Transform, Detail Preserving, Directional Decomposition, Bitplane Coding
ABSTRACT:
It is well known that the commonly used separable extensions of one-dimensional transforms, such as the wavelet transform, have
limitations in capturing the geometry of image edges. But natural scene images always have geographic lines, such as textures, edges
which can not be well reconstructed when compressed by using one-dimensional transform. So in lossy compression of images, it is
important to find a method which can best preserve the image by using a small description. To solve this problem, in this paper, we
present a new compression method which can preserve one of the most important cues of image: the directional details. In order to
show the ability of detail preserving lossy coder we present tests using typical images with much directional information, and a
comparison between wavelets and the new wavelet based contourlet transform(WBCT) is made. Finally, we present the results
obtained by using image quality assessment method. Our experiments demonstrate that the decompressed images based on WBCT
can preserve more directional information than wavelet transform, which can be used efficiently in future image processing such as
classification, edge detecting and so on.
1. INTRODUCTION
In recent years, a considerable effort have been made to design
image compression method in which the main goal is to obtain
good quality of decompressed images even at very low bit
rates(0. O. Vergara Villegas, 2006; K. R. Namuduri and V. N.
Ramaswamy, 2003). Due to the great use of digital information,
image compression becomes imperative in different areas such
as image storage, transmission and processing. At these areas
the representation of the information needs to be efficient. The
goal of image coding is to reduce the bit rate for signal
transmission or storage while maintaining an acceptable image
quality for different purposes (D. Schilling and P. C. Cosman,
2003).
Research of biologists on human vision system and natural
scene images static model shows that: the best image
representation should have characters below (Minh N.Do, and
Martin Vetterli, 2005):
1) Multiresolution. The representation should allow images to
be successively approximated, from coarse to fine resolutions.
2) Localization. The basis elements in the representation
should be localized in both the spatial and the frequency
domains.
3) Directionality. The representation should contain basis
elements oriented at a variety of directions, much more than the
few directions that are offered by separable wavelets.
A digital image is mainly composed by: edges, edge associated
details and textures, and this three parts are very important in
reconstructing an image. So in image compression, it is very
necessary to preserve this information to get a good quality of
reconstructed image. How to save these parts of an natural
scene image by using a small description is a question that
considered by many researchers.
Imagine that there are two painters, one with a “wavelet”-style
and the other with a new style, both wishing to paint a natural
scene. Both painters apply a refinement technique to increase
resolution from coarse to fine. Here, efficiency is measured by
how quickly, that is with how few brush strokes, and one can
faithfully reproduce the scene, which is shown below in Figure
1 (Minh N.Do, and Martin Vetterli, 2005).
Figure 1. Wavelet versus new scheme: illustrating the
successive refinement by the two systems
near a smooth contour, which is shown as a
thick curve separating two smooth regions?
When a smooth contour is being painted, as shown in Figure 1,
we can clearly see the limitation of the wavelet-style painter
who needs many fine dots to capture the contour, while the new
scheme can make different elongated shapes and in a variety of
directions following the smoothness of the contour. This decide
that wavelet transform can not take advantage of the geographic
character of natural images, and it is not the best and the most
sparse representation method of data. So a new function
1 This work was supported by China National “863” Foundation. No. 2006AA12Z136 and the National Natural Foundation of China
under Grant 40601055.