Choose some image, convert it to grayscale.
Use dwt2, choose mode (e.g. 'zpd') and wavelet type (e.g. 'haar').
It returns 4 matrices, visualize all of them.
Use formula from lecture to quantize 4 results.
Choose some value of delta. (e.g. 10).
Multiply the results with delta and call idwt2 .
Use wavedec2. (From documentation: output decomposition structure consists of the wavelet decomposition vector C and the bookkeeping matrix S )
Perform quantization just as you did in the first part - on output C, then run inverse steps as well (use waverec2).
To extract individual level - use wrcoef2. Type parameter corresponds to following:
'a' - approximation coefficients
'V' - vertical detail coefficients
'H' - horizontal detail coefficients
'D' - diagonal detail coefficients