No-reference quality assessment using natural scene statistics: JPEG2000

被引:398
作者
Sheikh, HR
Bovik, AC
Cormack, L
机构
[1] Univ Texas, Lab Image & Video Engn, Dept Elect & Comp Engn, Austin, TX 78712 USA
[2] Univ Texas, Dept Psychol, Austin, TX 78712 USA
基金
美国国家科学基金会;
关键词
blind quality assessment (QA); image QA; JPEG2000; natural scene statistics (NSS); no reference (NR) image QA;
D O I
10.1109/TIP.2005.854492
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Measurement of image or video quality is crucial for many image-processing algorithms, such as acquisition, compression, restoration, enhancement, and reproduction. Traditionally, image quality assessment (QA) algorithms interpret image quality as similarity with a "reference" or "perfect" image. The obvious limitation of this approach is that the reference image or video may not be available to the QA algorithm. The field of blind, or no-reference, QA, in which image quality is predicted without the reference image or video, has been largely unexplored, with algorithms focussing mostly on measuring the blocking artifacts. Emerging image and video compression technologies can avoid the dreaded blocking artifact by using various mechanisms, but they introduce other types of distortions, specifically blurring and ringing. In this paper, we propose to use natural scene statistics (NSS) to blindly measure the quality of images compressed by JPEG2000 (or any other wavelet based) image coder. We claim that natural scenes contain nonlinear dependencies that are disturbed by the compression process, and that this disturbance can be quantified and related to human perceptions of quality. We train and test our algorithm with data from human subjects, and show that reasonably comprehensive NSS models can help us in making blind, but accurate, predictions of quality. Our algorithm performs close to the limit imposed on useful prediction by the variability between human subjects.
引用
收藏
页码:1918 / 1927
页数:10
相关论文
共 40 条
[1]  
Bovik AC, 2001, INT CONF ACOUST SPEE, P1725, DOI 10.1109/ICASSP.2001.941272
[2]   Image compression via joint statistical characterization in the wavelet domain [J].
Buccigrossi, RW ;
Simoncelli, EP .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 1999, 8 (12) :1688-1701
[3]  
Cho S, 2002, IEEE T CIRC SYST VID, V12, P157
[4]   Multiscale image segmentation using wavelet-domain hidden Markov models [J].
Choi, H ;
Baraniuk, RG .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2001, 10 (09) :1309-1321
[5]   Can recent innovations in harmonic analysis 'explain' key findings in natural image statistics? [J].
Donoho, DL ;
Flesia, AG .
NETWORK-COMPUTATION IN NEURAL SYSTEMS, 2001, 12 (03) :371-393
[6]   RELATIONS BETWEEN THE STATISTICS OF NATURAL IMAGES AND THE RESPONSE PROPERTIES OF CORTICAL-CELLS [J].
FIELD, DJ .
JOURNAL OF THE OPTICAL SOCIETY OF AMERICA A-OPTICS IMAGE SCIENCE AND VISION, 1987, 4 (12) :2379-2394
[7]   THE PRINCIPAL COMPONENTS OF NATURAL IMAGES [J].
HANCOCK, PJB ;
BADDELEY, RJ ;
SMITH, LS .
NETWORK-COMPUTATION IN NEURAL SYSTEMS, 1992, 3 (01) :61-70
[8]  
Janssen TJWM, 2000, J IMAGING SCI TECHN, V44, P93
[9]   An embedded wavelet video coder using three-dimensional set partitioning in hierarchical trees (SPIHT) [J].
Kim, BJ ;
Pearlman, WA .
DCC '97 : DATA COMPRESSION CONFERENCE, PROCEEDINGS, 1997, :251-260
[10]   A mathematical analysis of the DCT coefficient distributions for images [J].
Lam, EY ;
Goodman, JW .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2000, 9 (10) :1661-1666