For Lenna (B/W) images, use 6, 8, 10 tap $h_o(n)$ digital FIR filters as given below;
0.0352 -0.0106 0.0033
-0.0854 0.0329 -0.0126
-0.1350 0.0308 -0.0062
0.4599 -0.1870 0.0776
0.8069 -0.0280 -0.0322
0.3327 0.6309 -0.2423
0.7148 0.1384
0.2304 0.7243
0.6038
0.1601
2- Implement Multiresolution Pyramid algorithm for given any 2-D B/W images. Algorithm
schema is shown below.
a) Show that it satisfies PR-Condition, and plot magnitude square frequency response of $h_o(n)$
$H_o(e^{j\omega})H_o(e^{-j\omega}) + H_1(e^{j\omega})H_1(e^{-j\omega}) = 2$
b) Obtain 4 iteration (13 Sub-band dyadic decomposition) for given FIR digital filters
$h_o(n)$ where N = 6,8,10 tap. And Re-construct the image by using the highest energy 5
sub-bands, and calculate the Compression Ratio and Peak Signal to Noise Ratio.
$\text{PSNR} = 10 \log_{10} \frac{255^2}{\sum \sum (x_i(r,c) - \hat{x}_i(r,c))^2}$
$h_o[n] = \delta[n] - \delta[n-1]$
$h_1[n] = \delta[n] + \delta[n-1]$
$g_o[n] = \delta[n] + \delta[n-1]$
$g_1[n] = \delta[n] - \delta[n-1]$
