Transcript Quality Assessment of Deblocked Images

Quality Assessment of Deblocked Images
C H A N G H O ON Y I M , M E M B E R , I E E E , A N D
ALAN CONRAD BOVIK, FELLOW, IEEE
Outline
 Introduction
 Quality assessment methods
 Simulation Results
 Concluding Remarks
INTRODUCTION
 Blocking effects are common in block-based image
and video compression systems. Deblocking filter
can improve image quality in some aspects, but can
reduce image quality in other regards.
 In this paper, we will review the image quality
assessment methods, and present the simulation
results on quality assessment of deblocked images
and videos.
 It also propose a new deblocking quality index,
PSNR-B.
Outline
 Introduction
 Quality assessment methods
 Simulation Results
 Concluding Remarks
QUALITY ASSESSMENT

PSNR (Peak Signal-to-Noise Ratio)
MSE 𝑥, 𝑦 =
1
𝑁
PSNR 𝑥, 𝑦 =
2552
10𝑙𝑜𝑔10
𝑀𝑆𝐸 𝑥,𝑦
𝑁
2
𝑖=1 𝑒𝑖
=
1
𝑁
𝑁
𝑖=1
𝑥𝑖 − 𝑦𝑖
2
QUALITY ASSESSMENT

SSIM (Structural Similarity)
Luminance comparison function:
l 𝑥, 𝑦 =
2𝜇𝑥 𝜇𝑦 +𝐶1
𝜇𝑥 2 +𝜇𝑦 2 +𝐶1
Contrast comparison function:
c 𝑥, 𝑦 =
2𝜎𝑥 𝜎𝑦 +𝐶2
𝜎𝑥 2 +𝜎𝑦 2 +𝐶2
Structure comparison function:
s 𝑥, 𝑦 =
𝜎𝑥𝑦 +𝐶3
𝜎𝑥 𝜎𝑦 +𝐶3
QUALITY ASSESSMENT

SSIM
SSIM 𝑥, 𝑦 = l 𝑥, 𝑦 ．c 𝑥, 𝑦 ．s 𝑥, 𝑦
=
2𝜇𝑥 𝜇𝑦 +𝐶1 2𝜎𝑥𝑦 +𝐶2
𝜇𝑥 2 +𝜇𝑦 2 +𝐶1 𝜎𝑥 2 +𝜎𝑦 2 +𝐶2
QUALITY ASSESSMENT

PSNR-B (PSNR Including Blocking Effects)
BEF 𝑦 =η． 𝐷𝐵 𝑦 − 𝐷𝐵𝐶 𝑦
η=
𝑙𝑜𝑔2 𝐵
𝑙𝑜𝑔2 𝑚𝑖𝑛 𝑁𝐻, 𝑁𝑉
=0
, if 𝐷𝐵 𝑦 > 𝐷𝐵𝐶 𝑦
, otherwise
𝐵𝐸𝐹𝑘 𝑦 = η． 𝐷𝐵𝑘 𝑦 − 𝐷𝐵𝐶𝑘 𝑦
𝐾
𝐵𝐸𝐹𝑇𝑜𝑡 𝑦 =
𝐵𝐸𝐹𝑘 𝑦
𝑘=1
QUALITY ASSESSMENT

PSNR-B (PSNR Including Blocking Effects)
MSE-B 𝑥, 𝑦 = MSE 𝑥, 𝑦 + 𝐵𝐸𝐹𝑇𝑜𝑡 𝑦
PSNR-B 𝑥, 𝑦 =
2552
10𝑙𝑜𝑔10
𝑀𝑆𝐸−𝐵 𝑥,𝑦
Outline
 Introduction
 Quality assessment methods
 Simulation Results
 Concluding Remarks
Simulation Results
Lena
Babara
Peppers
Simulation Results
PSNR comparison
of images.
(a) Lena
(b) Peppers
(c) Babara
(d) Goldhill
Simulation Results
SSIM comparison
of images.
(a) Lena
(b) Peppers
(c) Babara
(d) Goldhill
Simulation Results
PSNR-B comparison
of images.
(a) Lena
(b) Peppers
(c) Babara
(d) Goldhill
Simulation Results
Study of H.264 In-loop filter
Forema
n
Mother and
Daughter
Simulation Results
PSNR comparison
of filters for H.264
videos.
(a) Foreman
(b) Mother
(c) Hall Monitor
(d) Mobile
Simulation Results
SSIM comparison
of filters for H.264
videos.
(a) Foreman
(b) Mother
(c) Hall Monitor
(d) Mobile
Simulation Results
PSNR-B comparison
of filters for H.264
videos.
(a) Foreman
(b) Mother
(c) Hall Monitor
(d) Mobile
Outline
 Introduction
 Quality assessment methods
 Simulation Results
 Concluding Remarks
Concluding Remarks
 PSNR-B modifies the conventional PSNR by including an
effective blocking effect factor. The simulation results
show that PSNR-B results in better performance than
PSNR for image quality assessment of these impaired
images.
 Quality studies of this type using special-purpose quality
indices (such as PSNR-B) and perceptually proven
indices (such SSIM) in conjunction are of considerable
value, not only for studying deblocking operations, but
also for other image improvement applications, such as
restoration, denoising, enhancement, and so on.