No-reference peak signal to noise ratio estimation based on generalized Gaussian modeling of transform coefficient distributions

Ji-Woo Ryu; Seon-Oh Lee; Dong-Gyu Sim; Jong-Ki Han

doi:10.1117/1.OE.51.2.027401

29 February 2012 No-reference peak signal to noise ratio estimation based on generalized Gaussian modeling of transform coefficient distributions

Ji-Woo Ryu, Seon-Oh Lee, Dong-Gyu Sim, Jong-Ki Han

Author Affiliations +

Funded by: MKE/KEIT, IT R&D program of MKE/KEIT, National IT Industry Promotion Agency

Optical Engineering, Vol. 51, Issue 2, 027401 (February 2012). https://doi.org/10.1117/1.OE.51.2.027401

Abstract

We present a no-reference peak signal to noise ratio (PSNR) estimation algorithm based on discrete cosine transform (DCT) coefficient distributions from H.264/MPEG-4 part 10 advanced video codec (H.264/AVC) bitstreams. To estimate the PSNR of a compressed picture without the original picture on the decoder side, it is important to model the distribution of transform coefficients obtained from quantized coefficients accurately. Whereas several conventional algorithms use the Laplacian or Cauchy distribution to model the DCT coefficient distribution, the proposed algorithm uses a generalized Gaussian distribution. Pearson's χ² (chi-square) test was applied to show that the generalized Gaussian distribution is more appropriate than the other models for modeling the transform coefficients. The χ² test was also used to find optimum parameters for the generalized Gaussian model. It was found that the generalized Gaussian model improves the accuracy of the DCT coefficient distribution, thus reducing the mean squared error between the real and the estimated PSNR.

Citation Download Citation

Ji-Woo Ryu, Seon-Oh Lee, Dong-Gyu Sim, and Jong-Ki Han "No-reference peak signal to noise ratio estimation based on generalized Gaussian modeling of transform coefficient distributions," Optical Engineering 51(2), 027401 (29 February 2012). https://doi.org/10.1117/1.OE.51.2.027401

Published: 29 February 2012

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE