Type of Document Thesis Author Chaganti, Venkata Ravikiran URN etd-04192005-041139 Title Edge Detection of Noisy Images using 2-d Discrete Wavelet Transform Degree Master of Science Department Electrical and Computer Engineering, Department of Advisory Committee
Advisor Name Title Simon Foo Committee Chair Keywords
- Edge Detection
Date of Defense 2005-04-11 Availability unrestricted AbstractWavelets are mathematical functions that cut up data into different frequency components, and then study each component with a resolution matched to its scale. Wavelets are an extremely useful tool for coding images and other real signals. Because the wavelet transform is local in both time (space) and frequency, it localizes information very well compared to other transforms. Wavelets code transient phenomena, such as edges, efficiently, localizing them typically to just a few coefficients.
This thesis deals with the different types of edge detection techniques, mainly concentrating on the two major categories Gradient and Laplacian. The gradient method detects the edges by looking for the maximum and minimum in the first derivative of the image. The Laplacian method searches for zerocrossings in the second derivative of the image to find edges.
Given the wavelet transforms values wavelet analysis can be done in the wavelet domain by comparison of wavelet coefficients that account for the edges. The detection of the maxima or inflection points is generally a key factor for analyzing the characteristics of the non-stationary signals. The wavelet transformation has been proved to be a very promising technique for the multiscale edge detection applied both to 1-D and 2-D signals. The dyadic wavelet transforms at two adjacent scales are multiplied as a product function to magnify the edge structures and suppress the noise. Unlike many multiscale techniques that first form the edge maps at several scales and then synthesize them together, we determined the edges as the local maxima directly in the scale product after an efficient thresholding. It is shown that the scale multiplication achieves better results than either of the two scales, especially on the localization performance.
The thesis deals with the comparison of edge detection of images using traditional edge detection operators (Prewitt, Sobel, Frei-chen and Laplacian of Gaussian) and Discrete Wavelet Transformation (DWT) using Haar, Daubechies, Coifman and Biorthogonal wavelets. It also deals with the edge detection of noisy images and the optimization of the wavelets for edge detection.
Filename Size Approximate Download Time (Hours:Minutes:Seconds)
28.8 Modem 56K Modem ISDN (64 Kb) ISDN (128 Kb) Higher-speed Access Abstract_page.pdf 14.67 Kb 00:00:04 00:00:02 00:00:01 < 00:00:01 < 00:00:01 Acknowledgement_page.pdf 9.82 Kb 00:00:02 00:00:01 00:00:01 < 00:00:01 < 00:00:01 BIOGRAPHICAL_SKETCH.pdf 12.59 Kb 00:00:03 00:00:01 00:00:01 < 00:00:01 < 00:00:01 CHAPTER_1.pdf 501.26 Kb 00:02:19 00:01:11 00:01:02 00:00:31 00:00:02 CHAPTER_2.pdf 250.11 Kb 00:01:09 00:00:35 00:00:31 00:00:15 00:00:01 CHAPTER_3.pdf 78.06 Kb 00:00:21 00:00:11 00:00:09 00:00:04 < 00:00:01 CHAPTER_4.pdf 279.33 Kb 00:01:17 00:00:39 00:00:34 00:00:17 00:00:01 CHAPTER_5.pdf 2.75 Mb 00:12:43 00:06:32 00:05:43 00:02:51 00:00:14 CHAPTER_6.pdf 15.58 Kb 00:00:04 00:00:02 00:00:01 < 00:00:01 < 00:00:01 Dedication_page.pdf 9.49 Kb 00:00:02 00:00:01 00:00:01 < 00:00:01 < 00:00:01 List_of_figures.pdf 27.87 Kb 00:00:07 00:00:03 00:00:03 00:00:01 < 00:00:01 REFERENCES.pdf 31.59 Kb 00:00:08 00:00:04 00:00:03 00:00:01 < 00:00:01 Signature_page.pdf 10.53 Kb 00:00:02 00:00:01 00:00:01 < 00:00:01 < 00:00:01 TABLE_OF_CONTENTS.pdf 23.98 Kb 00:00:06 00:00:03 00:00:02 00:00:01 < 00:00:01 Title_page.pdf 11.25 Kb 00:00:03 00:00:01 00:00:01 < 00:00:01 < 00:00:01