INVERSE HALFTONING WITH GROUPING SINGULAR VALUE DECOMPOSITION
Jun Yang
1,2
Jun Guo
1
Hongyang Chao
3,∗
1
School of Information Science and Technology, Sun Yat-sen University, Guangzhou, P.R. China
2
SYSU-CMU Shunde International Joint Research Institute, P.R. China
3
School of Software, Sun Yat-sen University, Guangzhou, P.R. China
ABSTRACT
The objective of inverse halftoning refers to reconstruct a
high quality gray scale image from bi-level halftone im-
age. However, reconstructing continuous-tone images from
their halftoned versions is highly underdetermined, making
this technique very difficult. In this paper, we present the
Grouping Singular Value Decomposition (G-SVD), a novel
approach which first groups similar image patches as input
and then characterizes lower-dimensional regions in input s-
pace where the data density is peaked. By adding a constraint
formulated via G-SVD into inverse halftoning, noises are
separated from meaningful contents and similarity of nonlo-
cal image patches is promoted. Our experiments shown that
the proposed approach could improve the visual quality of
reconstructed results and outperformed the state of the arts in
terms of both objective and subjective measurements.
Index Terms— inverse halftoning, grouping singular val-
ue decomposition (G-SVD), nonlocal similarity
1. INTRODUCTION
Halftoning provides the ability of representing an image with
only one color through the use of ink dots, and is thus widely
used in today’s publishing applications, such as newspapers,
books, magazines, etc. [1, 2]. Halftone images are typically d-
ifficult to manipulate. Many halftone image processing, such
as scaling, compression, and enhancement could cause severe
image degradation [3]. To enable these operations, gray im-
ages need to be reconstructed from halftones through inverse
halftoning. For scanning printed (halftone) images inverse
halftoning is also needed to reduce Moire effects.
Therefore, as shown in Fig. 1, halftoned images are of-
ten inverse halftoned. Considering that inverse halftoning is
an operation mapping {0, 1}
H×W
onto R
H×W
(H / W is
the image height / width), this technique apparently belongs
to the class of ill-posed inverse problems and is rather chal-
lenging. In the last two decades, lots of inverse halftoning
This work was partially supported by NSF of China under Grant
61173081 and Guangdong Natural Science Foundation, China, under Grant
S2011020001215.
∗
Corresponding author: Hongyang Chao (isschhy@mail.sysu.edu.cn)
Fig. 1. An illustration of inverse halftoning: original image
(left), halftoned (center) and inverse halftoned version (right).
algorithms were developed to exploit local smoothness of im-
ages and have achieved good results, including tree-structured
vector quantization [4], convex set projection [5], iterative s-
tatistical smoothing [6], nonlinear permutation filtering [7],
anisotropic filtering [8], shearlet representation [9], Bayesian
approaches [10, 11], look-up-table (LUT) based approach-
es [12] and wavelet-based approaches [13, 14, 15]. However,
while local smoothness was preserved, these algorithms usu-
ally produced blurry results, and thus had shortcomings in
keeping edges and textures. Targeting at this problem, the re-
cently proposed BM3D-based approach [16] added a regular-
ization term of nonlocal similarity, and achieved the state-of-
the-art performance. However, there is still one main problem
to restrict its performance. The BM3D-based approach relied
on a sophisticated deterministic annealing optimization [17],
which requires to denoise reconstructed images in each itera-
tion. Hence, its output usually consists of visible artifacts.
To resolve this issue, while minimizing residual errors be-
tween an input halftoned image and the re-halftoned version
of its reconstructed image so as to maintain local smooth-
ness, we plug in a novel constraint expressed via the pro-
posed Grouping Singular Value Decomposition (G-SVD) ap-
proach to promote self-similarity of nonlocal image patches.
By grouping similar nonlocal image patches and then limiting
the minimal nonzero singular value of each group to be a large
number, our G-SVD constraint can capture a low-dimensional
manifold from each group. As a result, saliency patterns with-
in a group, e.g., the outline of a human’s face, can be dis-
covered, while at the same time artifacts are well separat-
ed, which in turn enhances the similarity of nonlocal patches
within this group. To optimize our whole framework, we al-
so leverage G-SVD to build an iterative projection method, in
order to ensure the lower bound of singular values of groups.
Extensive experiments show the superior performance of our