19-03-2012, 01:31 PM
NO-REFERENCE PERCEPTUAL QUALITY ASSESSMENT OF JPEG COMPRESSED IMAGES
10.1.1.4.4014.pdf (Size: 1.18 MB / Downloads: 34)
ABSTRACT
Human observers can easily assess the quality of a distorted
image without examining the original image as a reference.
By contrast, designing objective No-Reference (NR) quality
measurement algorithms is a very difficult task. Currently,
NR quality assessment is feasible only when prior knowledge
about the types of image distortion is available
INTRODUCTION
In recent years, there has been an increasing need to develop
objective measurement techniques that can predict image
/video quality automatically. Such methods can have various
applications. First, they can be used to monitor image/
video quality for quality control systems. Second, they
can be employed to benchmark image/video processing systems
and algorithms. Third, they can also be embedded
into image/video processing systems to optimize algorithms
and parameter settings. The most widely used objective image
quality/distortion metrics are Peak Signal-to-Nose Ratio
(PSNR) and Mean Squared Error (MSE), but they are
widely criticized as well for not correlating well with perceived
quality measurement.
OBJECTIVE NR QUALITY ASSESSMENT
JPEG is a block DCT-based lossy image coding technique.
It is lossy because of the quantization operation applied to
the DCT coefficients in each 8£8 coding block. Both blurring
and blocking artifacts may be created during quantization.
The blurring effect is mainly due to the loss of high
frequency DCT coefficients, which smoothes the image signal
within each block. Blocking effect occurs due to the discontinuity
at block boundaries, which is generated because
the quantization in JPEG is block-based and the blocks are
quantized independently.
CONCLUSIONS
We demonstrate a novel NR perceptual quality assessment
scheme for JPEG compressed images. Subjective experiments
were conducted to evaluate the quality of JPEG compressed
images. The features described in the paper effectively
capture the artifacts introduced by JPEG, and the nonlinear
fitting gives good agreement with MOS scores.
The method is computationally efficient since no complicated
transforms are computed and the algorithm can be
implemented without storing the entire image (or even a row
of pixels) in memory, which makes embedded implementations
easier. The basic methodology of the proposed method
can also been used to develop NR quality assessment methods
for H.26x/MPEG compressed video.