Computational Signal Processing with Wavelets
A. Teolis, Naval Research Laboratory, Washington D.C.
0-8176-3909-8 * 1997-Spring * $59.95. * Hardcover * 330 pages
|
|
Computational Signal Processing with Wavelets |
Back To Main Additional Resources Page
Image Processing
Here is an example of wavelet image processing using the 2D fast wavelet transform to clean up a noisy source image. In this example, the predefined image "clown" is used with a SNR=10dB. The result of loading the clown image with this SNR is depicted below.
After selecting the fwt2 in Transform menu and setting the number of decomposition levels (via the alpha parameter) to 3, depressing the Xform button causes the 2D fast wavelet transform to be computed (using the default Daub3 wavelet). A new window is spawned which depicts the result as shown below.
Setting a threshold value with the slider as shown above depressing the Thresh button to perform the thresholding and then depressing the inverse discrete wavelet transform IDWT button causes a new window to spawn that contains the original and inverse transform images side by side. For the case of the clown image the original (left) and reconstructed (right) are shown below. This processing shows how one may use the discrete wavelet transform to perform noise suppression since the processed image appears less noisy than the original.
675 Massachusetts Avenue, Cambridge, MA 01239 - (617) 876-2333 - Information - Ordering - Books - Top of Page
© 1997 - Birkhauser Boston
Please direct questions regarding this site to webmaster@birkhauser.com
