Math 336: Image Processing - Laboratory #5

1. Warm Up Exercise: Image Pyramid

Read in the image of Mandrill and construct an image pyramind using the provided MATLAB functions ipyramid.m and showPyramid.m.

Use the functions in the lptoolbox to construct the deepest laplacian pyramid possible for the image of Mandrill. Plot an image histogram for each level.

Add noise to an image with imnoise or another method. Examine the file tvdecode.m and use it to denoise the image.

warmup.m

2. Image Fusion and Denoising

Carry out the image fusion method given in the assignment handout on imageA and imageB.

fusion.m

3. Homework: Wavelet Denoising

Develop a MATLAB function for computing the J-level, J = 1, 2, ... , log2 N, Harr wavelet transform of an N x N image (assuming N is a power of 2). Your function can repeatedly call the function harr.m or you can start from scratch. Your function should take two arguments: an input image and the number of levels J you wish to compute. Your function should output an array of N x N wavelet coefficients.

hwt.m

Develop a MATLAB function for computing the inverse J-level, J = 1, 2, ... , log2 N, Harr wavelet transform of an N x N array of Harr wavelet coefficients. Your function could repeatedly call the function iharr.m which reverses the process of haar.m. Your function should take two arguments: an input array of Harr wavelet coefficients and the number of levels of J in that array. Your function should output an image reconstructed from the input coefficients.

ihwt.m

Test your forward and inverse Harr transform code by applying it to a grayscale image of your choosing. Add a small amount of Gaussian noise to your chosen image. Compute the harr transform. Set all coefficients to zero except those whose magnitude is larger than 3 sigma and reconstruct an estimate of the original image by applying the corresponding inverse transform. Display the result.

wavelet.m