01.01.2015 Views

Fractal Geometry in Image Processing

Fractal Geometry in Image Processing

Fractal Geometry in Image Processing

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

characteristic feature of nonl<strong>in</strong>ear (as opposed to l<strong>in</strong>ear)<br />

process is the <strong>in</strong>teraction (coupl<strong>in</strong>g) of different modes,<br />

which may lead to non-random signal phase structure. Such<br />

collective phase properties of the signal cannot be detected<br />

by l<strong>in</strong>ear spectral methods. <strong>Fractal</strong> dimension can be an<br />

useful measure for the characterization of<br />

electrophysiological time series (Klonowski, 2000).<br />

b. 2-D fractals and image analysis<br />

A digitized image is a pattern stored as a<br />

rectangular data matrix. It is dist<strong>in</strong>guished between b<strong>in</strong>ary<br />

images, grayscale images and color images. The ultimate<br />

goal of image analysis is the identification of a scene and all<br />

objects <strong>in</strong> the image. <strong>Image</strong> analysis can be described as a<br />

set of techniques required to extract symbolic <strong>in</strong>formation<br />

from the image data. Different techniques are available for<br />

perform<strong>in</strong>g image analysis, of which fractal methods are<br />

promis<strong>in</strong>g <strong>in</strong> certa<strong>in</strong> applications. The images can be studied<br />

by compar<strong>in</strong>g the fractal dimensions of the orig<strong>in</strong>al and<br />

transformed images. Another approach which has been put<br />

forth is the local fractal operator method. In this method, the<br />

fractal dimension correspond<strong>in</strong>g to each pixel of the image<br />

is computed. This is followed by segmentation technique to<br />

extract the region of <strong>in</strong>terest from the image.<br />

Marchette et al., (1997) have employed fractal<br />

based techniques <strong>in</strong> digital mammography to detect<br />

tumorous tissues. The tumor region was better identified<br />

when segmentation boundaries were <strong>in</strong>corporated <strong>in</strong>to the<br />

calculation. Similar techniques can be employed for the<br />

detection of an object from remotely sensed images or <strong>in</strong> the<br />

recognition of biological structures from plant images.<br />

1.6.3. <strong>Fractal</strong> techniques <strong>in</strong> classification<br />

IRACST- International Journal of Research <strong>in</strong> Management & Technology (IJRMT), ISSN: 2249-9563<br />

Vol. 2, No. 1, 2012<br />

the recognition of sk<strong>in</strong> samples to determ<strong>in</strong>e the age of a<br />

person. They are also used <strong>in</strong> f<strong>in</strong>gerpr<strong>in</strong>t recognition and<br />

fabrics process<strong>in</strong>g.<br />

1.7 Conclusion<br />

The fractal theory is still descriptive rather than <strong>in</strong>ferential.<br />

There are a lot of application for fractals <strong>in</strong> different fields.<br />

The basic concepts and applications of fractals are discussed<br />

<strong>in</strong> this chapter.<br />

References<br />

1. Philip, N.S., Wadadekar, Y., Kembhavi, A,, Joseph,<br />

K.B.,'A difference boost<strong>in</strong>g neural network for<br />

automated star-galaxy classification', Astronomy and<br />

Astrophysics, 385 : 1119-1126,2002.<br />

2. Polvere, M. and Nappi, M., 'Speed-Up In <strong>Image</strong><br />

Cod<strong>in</strong>g : Comparison of Methods', IEEE Trans, on<br />

image process<strong>in</strong>g 9, 2000.<br />

3. Revathy, K., and Nayar, S.R.P. '<strong>Fractal</strong> Analysis of<br />

Ionospheric Plasma Motion', <strong>Fractal</strong>s, 6 : 127- 130,<br />

1998.<br />

4. Revathy, K., Raju, G., Nayar, S.R.P., '<strong>Image</strong> Zoom<strong>in</strong>g<br />

by Wavelets', <strong>Fractal</strong>s, 8 : 247-253, 2000.<br />

5. Ruhl, M., Hartenste<strong>in</strong>, H. and Saupe, D., 'Adaptive<br />

Partition<strong>in</strong>g For <strong>Fractal</strong> <strong>Image</strong> Compression', IEEE<br />

International Conference on <strong>Image</strong> Process<strong>in</strong>g<br />

(ICIP'97), Santa Barbara, 1997.<br />

<strong>Fractal</strong> geometry is widely used <strong>in</strong> the study of<br />

image characteristics. For recognition of regions and objects<br />

<strong>in</strong> natural scenes, there is always need for features that are<br />

<strong>in</strong>variant and provide a good set of descriptive values for<br />

the region. There are many fractal features that can be<br />

generated from an image (Chaudhuri and Sarkar, 1995).<br />

The most commonly used fractal feature is the fractal<br />

dimension. In some applications, fractal dimension alone is<br />

capable of discrim<strong>in</strong>at<strong>in</strong>g one object from one another. But<br />

<strong>in</strong> certa<strong>in</strong> applications, fractal dimension alone may not be<br />

sufficient <strong>in</strong> identify<strong>in</strong>g the desired object. Here, a set of<br />

fractal features obta<strong>in</strong>ed on simple image transformations<br />

can be used as the feature set. Another technique is based<br />

on fractal signature. <strong>Fractal</strong> signature is designed us<strong>in</strong>g<br />

epsilon blanket method. The fractal signature assigns a<br />

unique signature to each sample. Samples belong<strong>in</strong>g to each<br />

class have similar signature which makes them dist<strong>in</strong>guish<br />

from one another easily.<br />

<strong>Fractal</strong> features are commonly used <strong>in</strong> textures.<br />

<strong>Image</strong>s that possess homogeneous regions can be easily<br />

classified us<strong>in</strong>g these features. <strong>Fractal</strong> features are used <strong>in</strong><br />

6. Saar, E., 'Towards fractal description of Structure',<br />

332. Morphological Cosmology Proceed<strong>in</strong>gs,, Cracon,<br />

Poland, 205-218, 1988.<br />

7. Saha, S. and Vemuri, R., 'An analysis on the effect of<br />

image features on Lossy Cod<strong>in</strong>g Performance', IEEE<br />

Signal Process<strong>in</strong>g Letters, 7 : 104-107,2000.<br />

8. Sarkar, N. and Chaudhuri, B., 'An efficient approach to<br />

estimate fractal dimension of textural images', Pattern<br />

Recognition, 25(9) : 1035-1041,1992.<br />

9. Saupe, D., 'Break<strong>in</strong>g the time-complexity of fractal<br />

image compression', Technical Report 53, Institut fur<br />

Informatik, Universitat Freiburg, 1994.<br />

10. Saupe, D. and Hamzaoui, R., 'A review of the fractal<br />

image compression literature', ACM Computer<br />

Graphics, 28 : 268 - 276,1994.<br />

11. Sato, T., Matsuoka, M. and Takayasu, H., '<strong>Fractal</strong><br />

113

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!