e178-04L6

# e178-04L6 - Quantization Optimal Quantizer Uniform...

This preview shows pages 1–7. Sign up to view the full content.

Quantization (Jan 22, 2004) & Optimal Quantizer & Uniform Quantizer

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Image Quantization u (continuous ) Quantizer u& ε {r 1 ,r 2 ,r 3 ,.... ,r L } u& u t 1 t L+1 t k t k+1 r k
Decision/Reconstruction Levels u ε [t k ,t k+1 ] r k {t k : k=1,2,. ...,L+1} Transition or decision levels r k k th reconstruction level Example: Uniform quantizer u ε [0,10.0] We want u& ε {0,1,. ....,255} t 1 = 0; t 257 = 10.0; uniformly spaced, t k = (k-1).10/256 k = 1,2,. ....,257)

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Example: quantization rt t qt t r r kk k k k k =+ F H G I K J =− = −− 1 2 10 256 5 256 11 Quantization interval Constant Uniform quantizer
MMSE Quantizer Minimise the mean squared error, MSE = Expected value of (u-u&) 2 given the number of quantization levels L. Assume that the density function p u (u) is known (or can be approximated by a normalised histogram). Note that for images, u==image intensity. p u (u) is the image intensity ditribution.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Optimum MSE quantizer Ε Ε () ,( ( ) ) ( ) ( ) ( ) ( ) [, ] , ( ) ; uu p u d u MSE u u u u p udu ur ut t p u d u rt u t kk k i t t i L u L i i = =− =∈ == + = z zz z + + Expected value of u = Since if
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### Page1 / 10

e178-04L6 - Quantization Optimal Quantizer Uniform...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online