Show all abstracts
View Session
- Front Matter: Volume 8292
- Vision and HDR I
- Vision and HDR II
- Color Management
- The Dark Side of the Color
- Image Processing I
- Image Processing II
- Applications
- Printing and Halftoning I
- Printing and Halftoning II
- Spectral and Display
- Poster Session
Front Matter: Volume 8292
Front Matter: Volume 8292
Show abstract
This PDF file contains the front matter associated with SPIE Proceedings Volume 8292, including the Title Page, Copyright information, Table of Contents, and the Conference Committee listing.
Vision and HDR I
Color assimilation and contrast near absolute threshold
Show abstract
Simultaneous Contrast and Assimilation test targets are almost always viewed at high light levels. We measured the
appearances of Simultaneous Contrast, Assimilation and other spatial surrounds near absolute rod threshold. Given the
very different spatial organizations of receptive fields in rod and cone vision at detection threshold, it is not obvious that
these familiar cone-vision spatial effects would be observed at rod light levels. Nevertheless, the spatial experiments
showed that these targets have the same changes in appearance as those observed in bright light. Our experiments used
very dim candle light that was above threshold for rods and L cones, and below threshold for M and S cones. Although
detection threshold experiments show very different spatial organizations for rod and cone vision, we found that spatial
contrast experiments gave the same changes of appearance. Neural contrast mechanisms at the lowest end of our visual
HDR range are very similar to those at the top of the range in sunlight. This is true for both chromatic and achromatic
targets.
Multi-illuminant color constancy for HDR images through exposure segmentation
Show abstract
A standard practice in high dynamic range imaging is to compose an image through exposure bracketing which captures
a series of exposures of the same scene and then combine them together, followed by dynamic rang compression and
some color processing steps. Scenes lit by multiple illuminants such as a room with an artificial light source when the
sun is shining through the window is an often encountered scenario which offers opportunity for the high dynamic range
feature of an image pipeline to show its advantages. Traditional color constancy algorithms estimate a global white point
of the scene and then apply color correction based on this estimate, which could exaggerate the difference between the
illuminants, making part of the image better and part of the image worse, or compromise the color of the whole scene.
In this paper, we propose a method for the color constancy of high dynamic range scenes with multiple illuminants
utilizing the inherent difference in their luminance levels to assist the segmentation of the image into differently
illuminated portions and apply their corresponding color constancy parameters. Experimental results using two
exposures show superior performance of the proposed algorithm compared to traditional algorithms applying global
corrections only.
Vision and HDR II
Tone mapping for HDR images with dimidiate luminance and spatial distributions of bright and dark regions
Show abstract
This paper proposes a novel tone mapping method in consideration of human's perception for a high dynamic range (HDR) image with dimidiated luminance and spatial distributions of bright and dark regions. In order to represent an HDR image with a low dynamic range (LDR) display, it is necessary to appropriately compress a dynamic range of HDR image by tone mapping.There are some HDR images which cannot represent the real scene precisely by applying conventional tone mapping methods. In this study, we view an HDR image with dimidiated luminance and spatial distributions of bright and dark regions as a target image for our work,we assume that human's perception dose not feel a sense of discomfort even if a magnitude relationship between luminance values of pixels near the boundary
of the regions is reversed, when bright and dark regions are definitely divided according to dimidiated luminance and spatial distributions. Under the assumption, we divide HDR image into bright and dark regions and apply a tone mapping method to each region independently. In experiments, we will show that our tone mapping method produces the image represented by utilizing a dynamic range effectively. In addition, we will confirm that our tone mapping method is useful through subjective evaluation and discuss the features of the HDR images which are supposed to be suitable for the proposed method.
Color universal design: analysis of color category dependency on color vision type (3)
Show abstract
We report on the results of a study investigating the color perception characteristics of people with red-green color
confusion. We believe that this is an important step towards achieving Color Universal Design. In Japan, approximately
5% of men and 0.2% of women have red-green confusion. The percentage for men is higher in Europe and the United
States; up to 8% in some countries. Red-green confusion involves a perception of colors different from normal color
vision. Colors are used as a means of disseminating clear information to people; however, it may be difficult to convey
the correct information to people who have red-green confusion. Consequently, colors should be chosen that minimize
accidents and that promote more effective communication. In a previous survey, we investigated color categories
common to each color vision type, trichromat (C-type color vision), protan (P-type color vision) and deuteran (D-type
color vision). In the present study, first, we conducted experiments in order to verify a previous survey of C-type color
vision and P-type color vision. Next, we investigated color difference levels within "CIE 1976 L*a*b*" (the CIELAB
uniform color space), where neither C-type nor P-type color vision causes accidents under certain conditions (rain
maps/contour line levels and graph color legend levels). As a result, we propose a common chromaticity of colors that
the two color vision types are able to categorize by means of color names common to C-type color vision. We also offer
a proposal to explain perception characteristics of color differences with normal color vision and red-green confusion
using the CIELAB uniform color space. This report is a follow-up to SPIE-IS & T / Vol. 7528 7528051-8 and SPIE-IS
& T /vol. 7866 78660J-1-8.
Colour perception with changes in levels of illumination
Show abstract
The perceived colour of a stimulus depends on the conditions under which it is viewed. For colours employed as an
important cue or identifier, such as signage and brand colours, colour reproduction tolerances are critically important.
Typically, such stimuli would be judged using a known level of illumination but, in the target environment, the level of
illumination used to view the samples may be entirely different. The effect of changes in the viewing condition on the
perceptibility and acceptability of small colour differences should be understood when such tolerances and associated
viewing conditions, are specified.
A series of psychophysical experiments was conducted to determine whether changes in illumination level significantly
alter acceptability and perceptibility thresholds of uniform colour stimuli. It was found that perceived colour
discrimination thresholds varied by up to 2.0 ΔE00. For the perceptual correlate of hue however, this value could be of
significance if the accepted error of colour difference was at the threshold, thereby yielding the possibility of rejection
with changes in illumination level. Lightness and chroma on the other hand, exhibited greater tolerance and were less
likely to be rejected with illuminance changes.
Color Management
Reducing the number of calibration patterns for the two-by-two dot centering model
Show abstract
The two-by-two dot centering model enables predicting the spectral reflectance of color halftones and does not
depend on a specific halftoning algorithm. It requires measuring the reflectances of a large number of two-by-two
calibration tile patterns. Spectral measurement of hundreds or thousands of tile patterns is cumbersome and time
consuming. In order to limit the number of measurements, we estimate the reflectances of a large majority of two-by-
two calibration tile patterns from a small subset comprising less than 10% of all tile patterns. Using this subset
of measured two-by-two calibration tile patterns, we perform a linear regression in the absorptance space and
derive a transformation matrix converting tile pattern colorant surface coverages to absorptances. This
transformation matrix enables calculating the absorptance of all remaining two-by-two tile patterns. For a cyan,
magenta and yellow print, with 72 two-by-two measured calibration tile patterns, we are able to create a two-by-two
dot centering model having an accuracy only slightly below the accuracy of the model with the fully measured
set of 1072 two-by-two tile patterns.
Spatial gamut mapping for preserving the details of an image
Show abstract
The spatial gamut-mapping algorithm (SGMA) overcomes the drawbacks of the widely used color-by-color methods.
Spatial gamut mapping can preserve detailed information in original images by performing adaptive gamut mapping in
surrounding pixels within the image. However, spatial gamut mapping can result in hue shift and the halo effect. In
addition, it only preserves the boundary information outside the color gamut; the resulting gamut-mapped image does not
sufficiently preserve the detailed information in the input image. In this paper, we propose an SGMA that utilizes details
of the input image. Our approach improves detail that is not effectively represented with conventional spatial gamut
mapping. This is done by taking an original image and first implementing gamut mapping of the input image. Then, the
details of the input image and gamut-mapped image are extracted. By examining the out-of-gamut region, the details of
the input image can be preserved when these values are added to the gamut-mapped image. The resulting image is
obtained by clipping out-of-gamut pixels, since these pixels are generated in the process of preserving details. We
demonstrated that images obtained using the proposed method are more similar to the input images, compared to images
obtained using conventional methods.
Evaluating color calibration kits with virtual display
Show abstract
Routine color calibration is imperative for medical applications that rely on color fidelity such as digital pathology,
endoscopy, and colposcopy. However, commercially available products vary greatly in price and performance with no
available evaluation standard. Related studies have used only one or a few displays to evaluate the performance of color
calibration kits. We propose the concept of Virtual Display, a universal display platform that emulates the colorimetric
response of real displays. A wide-color-gamut display driven by an FPGA is used to emulate the colorimetric response of
real display devices. By changing the look-up tables in the FPGA, the virtual display emulates various real displays for
testing different color calibration kits. Our experimental data for 6 real displays show that the virtual display can emulate
real displays reasonably well. The results demonstrate that the proposed virtual display approach is a fast, economical,
and objective method for evaluating the performance of color calibration kits.
Optimizing color fidelity in wide gamut display devices when processing images compressed by block-based discrete cosine transforms (DCT)
Show abstract
High-end monitors based on LCD technology increasingly address wide color gamut implementations featuring precise
color calibration within a variety of different color spaces such as extended sRGB or AdobeRGB. However, images are
often reconstructed from digitally compressed images files such as JPEG or MPEG where color quality could be
questionable. In particular, when such image files are scaled up or zoomed in, different types of image artifacts become
visually noticeable. Among these artifacts we find pixel repetition, blockiness, ringing, and color blotching. While pixel
repetition and ringing appear due to insufficient adaptation to image context using a static or context adaptive filter
kernel in temporal domain, blockiness and ringing occur due to image compression in frequency domain, when image
compression factors are significant. In addition, chrominance channels often undergo an even higher compression ratio
that amplifies visibility of artifacts such as color blotches. Consequently, we are interested in improving the quality of
images to be displayed depending on image zoom factors. We propose to discriminate most relevant visual artifacts
using power spectrum analysis in DCT domain as well as kernel based rescaling combined with statistical analysis
taking into account characteristic non-stationary behavior of image content and identifiable visual artifacts. A
comparative analysis based on some competitive solutions highlights the effectiveness of our approach and identifies its
current limitations with regard to wide color gamut representation, primarily due to mathematical uncertainty of the
studied artifacts.
Optimal gamut volume design for three primary and multiprimary display systems
Show abstract
Primary selection plays a fundamental role in display design. Primaries affect not only the gamut of colors the
systems is able to reproduce, but also, they have an impact on the power consumption and other cost related
variables. Using more than the traditional three primaries has been shown to be a versatile way of extending
the color gamut, widening the angle view of LCD screens and improving power consumption of displays systems.
Adequate selection of primaries requires a trade-off between the multiple benefits the system offers, the costs
and the complexity it implies, among other design parameters.
The purpose of this work is to present a methodology for optimal design for three primary and multiprimary
display systems. We consider the gamut in perceptual spaces, which offer the advantage of an evaluation that
correlates with human perception, and determine a design that maximize the gamut volume, constrained to
a certain power budget, and analyze the benefits of increasing number of primaries, and their effect on other
variables of performance like gamut coverage.
The Dark Side of the Color
The dark side of CIELAB
Show abstract
Standardized in 1976 as a uniform color space, CIELAB is extensively utilized in color science and engineering
applications. CIELAB provides both a color difference formula and correlates for common perceptual descriptors
of color. Deficiencies in both areas are well-known, and based on these known limitations, numerous fixes
have been developed yielding alternative color difference formulae that are derived as modifications of the color
difference in CIELAB. In addition, several new color appearance spaces have also been proposed as modifications
of the basic CIELAB framework.
In this paper, we point out other, lesser-known and poorly-appreciated, limitations of CIELAB that occur
particularly in the dark regions of color space. We demonstrate via examples, how these limitations not only cause
performance compromises but lead to fundamental breakdowns in system optimization and design problems,
making CIELAB unusable in these problems. We consider the reasons why these fundamental limitations were
overlooked in the original development of CIELAB and analyze the mathematical representations contributing
to the undesired behavior. We argue that fundamental new research is required to overcome this dark side of
CIELAB; the development of uniform color spaces and new color appearance spaces must be revisited afresh
using new experimental data and keeping in mind newer devices and applications.
Complexities of complex contrast
Show abstract
For the visual system, luminance contrast is a fundamental property of images, and is one of the main inputs of any
simulation of visual processing. Many models intended to evaluate visual properties such as image discriminability
compute perceived contrast by using contrast sensitivity functions derived from studies of human spatial vision. Such use
is of questionable validity even for such applications (i.e. full-reference image quality metrics), but it is usually
inappropriate for no-reference image quality measures. In this paper, we outline why the contrast sensitivity functions
commonly used are not appropriate in such applications, and why weighting suprathreshold contrasts by any sensitivity
function can be misleading. We propose that rather than weighting image contrasts (or contrast differences) by some
assumed sensitivity function, it would be more useful for most purposes requiring estimates of perceived contrast or
quality to develop an estimate of efficiency: how much of an image is making it past the relevant thresholds.
It's not the pixel count, you fool
Show abstract
The first thing a "marketing guy" asks the digital camera engineer is "how many pixels does it have, for we need as
many mega pixels as possible since the other guys are killing us with their "umpteen" mega pixel pocket sized digital
cameras. And so it goes until the pixels get smaller and smaller in order to inflate the pixel count in the never-ending
pixel-wars. These small pixels just are not very good. The truth of the matter is that the most important feature of digital
cameras in the last five years is the automatic motion control to stabilize the image on the sensor along with some very
sophisticated image processing. All the rest has been hype and some "cool" design. What is the future for digital
imaging and what will drive growth of camera sales (not counting the cell phone cameras which totally dominate the
market in terms of camera sales) and more importantly after sales profits? Well sit in on the Dark Side of Color and find
out what is being done to increase the after sales profits and don't be surprised if has been done long ago in some
basement lab of a photographic company and of course, before its time.
Dark texture in artworks
Show abstract
This presentation highlights issues relating to the digital capture printing of 2D and 3D artefacts and accurate colour
reproduction of 3D objects.
There are a range of opportunities and technologies for the scanning and printing of two-dimensional and threedimensional
artefacts [1]. A successful approach of Polynomial Texture Mapping (PTM) technique, to create a
Reflectance Transformation Image (RTI) [2-4] is being used for the conservation and heritage of artworks as these
methods are non invasive or non destructive of fragile artefacts. This approach captures surface detail of twodimensional
artworks using a multidimensional approach that by using a hemispherical dome comprising 64 lamps to
create an entire surface topography. The benefits of this approach are to provide a highly detailed visualization of the
surface of materials and objects.
Harmonious colors: from alchemy to science
Show abstract
There is a very long tradition in designing color palettes for various applications, going back to at least the Upanishad.
Although color palettes have been influenced by the available colorants, starting with the advent of aniline dyes in the late
1850s there have been few physical limits on the choice of individual colors. This abundance of choices exacerbates the
problem of limiting the number of colors in a palette, i.e., in keeping them into a manageable quantity.
For example, it is not practical for a car company to offer each model in hundreds of colors. Instead, for each model
year a small number of color palettes is offered, each containing the colors for the body, trim, interior, etc. Another example
is the fashion industry, where in addition to solid colors there are also patterns, leading to a huge variety of combinations
that would be impossible to stock.
The traditional solution is that of "color forecasting." Color consultants assess the sentiment or affective state of a
target customer class and compare it with new colorants offered by the industry. They assemble a limited color palette,
name the colors according to the sentiment, and publish their result. Textile manufacturers will produce fabrics in these
colors and fashion designers will design clothes, accessories, and furniture based on these fabrics. Eventually, the media
will communicate these forecasts to the consumers, who will be admired by their cohorts when they choose colors from
the forecast palette, which by then is widely diffused.
The color forecasting business is very labor intensive and difficult, thus for years computer engineers have tried to come
up with algorithms to design harmonious color palettes, alas with little commercial success. For example, Johannes Itten's
color theory has been implemented many times, but despite Itten's success in the Bauhaus artifacts, the computer tools
have been of little utility. Indeed, contrary to the auditory sense, there is no known physiological mechanism sustaining
harmony and the term "harmonious" just has the informal meaning of "going well together."
We argue that the intellectual flaw resides in the belief that a masterful individual can devise a "perfect methodology"
that the engineer can then reduce to practice in a computer program. We suggest that the correct approach is to consider
color forecasting as an act of distillation, where a palette is digested from the sentiment of a very large number of people.
We describe how this approach can be reduced to an algorithm by replacing the subjective process with a data analytic
process.
Image Processing I
Detection of backlight images using chrominance
Show abstract
An image taken under the backlight condition shows that a main object or foreground appears very dark, but
a background appears relatively bright since the exposure time of the main object or foreground is relatively
shorter than the one of the background due to high luminance from the background. The determination of
the backlight image is generally done by luminance histogram analysis since it is believed that the distinct
characteristic of the backlight image is a large luminance difference between the foreground and background.
However, this conventional detection method may not be adequate for video images since it generally targets
on still images. Furthermore, the detection of the backlight image would not be performed well if there are
abrupt changes in light, motion, or scenes. Inaccurate detection leads to unnecessary compensation that makes
images over-highlighted or flickered, especially when consecutive frames of video have different illumination
modes. Since an image taken under normal light conditions may also have the similar luminance characteristics
of the backlight image, it would not be sufficient to discriminate between the normal light and backlight image
using only luminance information. Therefore, the analysis of chrominance of images is introduced to detect the
backlight image more accurately.
A new method for skin color enhancement
Show abstract
Skin tone is the most important color category in memory colors. Reproducing it pleasingly is an important factor in
photographic color reproduction. Moving skin colors toward their preferred skin color center improves the skin color
preference on photographic color reproduction. Two key factors to successfully enhance skin colors are: a method to
detect original skin colors effectively even if they are shifted far away from the regular skin color region, and a method
to morph skin colors toward a preferred skin color region properly without introducing artifacts. A method for skin
color enhancement presented by the authors in the same conference last year applies a static skin color model for skin
color detection, which may miss to detect skin colors that are far away from regular skin tones. In this paper, a new
method using the combination of face detection and statistical skin color modeling is proposed to effectively detect skin
pixels and to enhance skin colors more effectively.
Image Processing II
CIE chromaticity, Planckian locus, and correlated color temperature estimation from raw image data using color checker training images
Show abstract
In this paper we discuss and quantitatively evaluate the mapping of raw sensor chromaticities, i.e., r = R/(R +
G+B) and b = B/(R+G+B), into the CIE 1931 xy chromaticity space, with the constraint that solely training
chromaticities are being used which have been obtained using a color checker at a certain illumination. The region
near the Planckian locus is considered to be most relevant and a least-squares weighting scheme is proposed to
minimize the residuals in this region. Furthermore, the Planckian and daylight loci are approximated in the rb
raw sensor chromaticity space using color checker chromaticities at three illuminations, those commonly available
in light-booths. The effect of daylight emulation compared to the standard daylight illumination is evaluated.
In another part of this paper the mapping of rb chromaticities to correlated color temperature is discussed and
evaluated. The proposed method is based on a weighted least-squares fit of a 2nd-order 2D polynomial and
outperforms two other estimation methods. We present a comprehensive set of simulation results with real
measurements of reflectance, sensitivity, and emission spectrums.
Comparative performance analysis of mobile displays
Show abstract
Cell-phone display performance (in terms of color quality and optical efficiency) has become a critical factor in creating
a positive user experience. As a result, there is a significant amount of effort by cell-phone OEMs to provide a more
competitive display solution. This effort is focused on using different display technologies (with significantly different
color characteristics) and more sophisticated display processors. In this paper, the results of a mobile-display
comparative performance analysis are presented. Three cell-phones from major OEMs are selected and their display
performances are measured and quantified. Comparative performance analysis is done using display characteristics such
as display color gamut size, RGB-channels crosstalk, RGB tone responses, gray tracking performance, color accuracy,
and optical efficiency.
Termites: a Retinex implementation based on a colony of agents
Show abstract
This paper describes a novel implementation of the Retinex algorithm with the exploration of the image done
by an ant swarm. In this case the purpose of the ant colony is not the optimization of some constraints but
is an alternative way to explore the image content as diffused as possible, with the possibility of tuning the
exploration parameters to the image content trying to better approach the Human Visual System behavior. For
this reason, we used "termites", instead of ants, to underline the idea of the eager exploration of the image.
The paper presents the spatial characteristics of locality and discusses differences in path exploration with other
Retinex implementations. Furthermore a psychophysical experiment has been carried out on eight images with
20 observers and results indicate that a termite swarm should investigate a particular region of an image to find
the local reference white.
A color quantization algorithm based on minimization of modified Lp norm error in a CIELAB space
Show abstract
Color quantization algorithms are used to select a small number of colors that can accurately represent the
content of a particular image. In this research, we introduce a novel color quantization algorithm which is based
on the minimization of a modified Lp norm rather than the more traditional L2 norm associated with mean square
error (MSE). We demonstrate that the Lp optimization approach has two advantages. First, it distributes the
colors more uniformly over the regions of the image; and second, the norm's value can be used as an effective
criterion for selecting the minimum number of colors necessary to achieve accurate representation of the image.
One potential disadvantage of the modified Lp norm criteria is that it could increase the computation of the
associated clustering methods. However, we solve this problem by introducing a two stage clustering procedure in
which the first stage (pre-clustering) agglomerates the full set of pixels into a relatively large number of discrete
colors; and the second stage (post-clustering) performs modified Lp norm minimization using the reduced number
of discrete colors resulting from the pre-clustering step. The number of groups used in the post-clustering is then
chosen to be the smallest number that achieves a selected threshold value of the normalized Lp norm. This two-stage
clustering process dramatically reduces computation by merging together colors before the computationally
expensive modified Lp norm minimization is applied.
Applications
Human skin imaging using three-phase spectral matching imager
Show abstract
A method for detecting human skin and identifying the surface condition of it using the three-phase spectral
matching imager (3PSMI) is proposed. The 3PSMI produces correlations pixel-wise between the spectrum of
an unknown object and an orthonormal pair of known reference spectra by use of the correlation image sensor
(CIS), and outputs the result as a complex image at an ordinary video frame rate, thus realizing both real-time
operation and high spectral resolution. In order to apply the 3PSMI to human skin, two tasks-discrimination
between skin and non-skin objects and evaluation of the degree of suntan and congestion on skin-are considered
and methods of generating an orthonormal pair of reference spectra for these tasks are proposed. Experiments
on these tasks are carried out on the developed 3PSMI with the reference spectra generated by the proposed
methods.
Appearance analysis of human skin with cosmetic foundation
Show abstract
We describe a method of analyzing the appearance of cosmetic foundation applied to the human face. In particular, we
focus on the "oily-shine" appearance, which is caused by sebum. A multi-band camera system with six spectral channels
is used for the analysis of the oily-shine appearance. As a basic analysis, we examine the optical features of oily-shine
by using two artificial skins looking like make-up skin with oily-shine and without oily-shine. We show that oily-shine
can be defined as the standard dichromatic reflection model. On the basis of the above findings, we propose a method
for detecting oily-shine area. This method involves (1) the extraction candidate areas, and (2) the evaluation of
appearance with oily-shine. First, we capture the CIE XYZ tri-stimulus image of an original make-up face by using the
multi-band camera and after a few hours later, capture the same face as a test facial image. Second, the candidate areas
with oily-shine are extracted by applying the Laplacian operator to luminance Y component of the test facial image.
Third, the principal component analysis is performed on the set of luminance and chromaticity (Y, x, y) of each candidate
area. Light reflection of oil-shine is regarded as the specular component of the dichromatic reflection. Finally, we
determine the existence of oily-shine by comparing specular clusters between the original image and the test image. The
proposed method is tested in experiments with subjective assessment for various real make-up facial images.
Color analysis and image rendering of woodblock prints with oil-based ink
Show abstract
This paper proposes a method for analyzing the color characteristics of woodblock prints having oil-based ink and
rendering realistic images based on camera data. The analysis results of woodblock prints show some characteristic
features in comparison with oil paintings: 1) A woodblock print can be divided into several cluster areas, each with
similar surface spectral reflectance; and 2) strong specular reflection from the influence of overlapping paints arises only
in specific cluster areas. By considering these properties, we develop an effective rendering algorithm by modifying our
previous algorithm for oil paintings. A set of surface spectral reflectances of a woodblock print is represented by using
only a small number of average surface spectral reflectances and the registered scaling coefficients, whereas the previous
algorithm for oil paintings required surface spectral reflectances of high dimension at all pixels. In the rendering process,
in order to reproduce the strong specular reflection in specific cluster areas, we use two sets of parameters in the
Torrance-Sparrow model for cluster areas with or without strong specular reflection. An experiment on a woodblock
printing with oil-based ink was performed to demonstrate the feasibility of the proposed method.
Printing and Halftoning I
Pre-RIP color management for soft proofing
Ingeborg Tastl,
Kok-Wei Koh
Show abstract
In an ideal situation, a color managed soft proof is generated by the same Raster Image Processor (RIP) that produces the
final raster data for printing. In real-world scenarios, this soft proofing capability is often absent. To overcome this
limitation, we developed a workflow using the Adobe PDF library for producing color managed soft proofs before
incoming jobs from the web are sent to the RIP. This paper discusses color management challenges that presented
themselves and how they were addressed. It also highlights a few specific cases where the separation into subsequent
color management and transparency blending processes caused unexpected results and how they were dealt with. Such a
system can be implemented as a cloud-based solution to provide end-customers with a quick, inexpensive and
reasonably accurate preview of print products before the job is submitted.
Parametrically controlled, stochastically seeded clustered halftones
Show abstract
We present a stochastically seeded halftoning method with parametric control of dot shapes as well as seed placement
adaption to local image structure. While the literature on other randomly structured screens (blue noise, green noise,
FM) is quite extensive, there is very little discussion on optimization of dot shape for robustness or preferred tile
appearance. The halftoning method that we introduce defines dot centers as seeds that are placed, typically in a random
high spatial frequency configuration. Spot functions are defined about these randomly placed seeds, where the spot
function allows control of dot cluster growth, touch points, cluster angle, and eccentricity. The spot function can also be
applied to regular and irregular polygonal halftone tiling. The seed adaption aspect of the halftoning method allows for
better edge rendition than conventional isotropic methods.
Assessing color reproduction tolerances in commercial print workflow
Show abstract
Except for linear devices like CRTs, color transformations from colorimetric specifications to device coordinates are mostly
obtained by measuring a set of samples, inverting the table, and looking up values in the table (including interpolation),
and mapping the gamut from input to output device.
The accuracy of a transformation is determined by reproducing a second set of samples and measuring the reproduction
errors. Accuracy as the average predicted perceptual error is then used as a metric for quality. Accuracy and precision are
important metrics in commercial print because a print service provider can charge a higher price for more accurate color,
or can widen his tolerances when customers prefer cheap prints.
The disadvantage of determining tolerances through averaging perceptual errors is that the colors in the sample sets
are independent and this is not necessarily a good correlate of print quality as determined through psychophysics studies.
Indeed, images consist of color palettes and the main quality factor is not color fidelity but color integrity. For example, if
the divergence of the field of error vectors is zero, color constancy is likely to take over and humans will perceive the color
reproduction as being of good quality, even if the average error is relatively large. However, if the errors are small but in
random directions, the perceived image quality is poor because the relation among colors is altered.
We propose a standard practice to determine tolerance based on the Farnsworth-Munsell 100-hue test (FM-100) for
the second set and to evaluate the color transpositions-a metric for color integrity-instead of the color differences. The
quality metric is then the FM-100 score. There are industry standards for the tolerances of color judges, and the same tolerances
and classification can be use for print workflows or its components (e.g., presses, proofers, displays). We generalize
this practice to arbitrary perceptually uniform scales tailored to specific applications and present an implementation.
In essence, we propose to extend the color discrimination test procedures used to evaluate human observers, to mechanical
and electronic color reproduction devices.
Investigating the paper dependency of laser printed colors for uncoated papers
S. Gorji Kandi
Show abstract
There are several parameters, which influence the characteristics of the final printed colors in digital printing. One of the
most important one is the substrate (paper). In the present study, the effect of uncoated papers on the laser printed colors
was investigated. For this purpose, three kinds of paper with the same grammage value of 80 and different whiteness and
texture were selected. The effect of paper on the printer's reproducibility, light-fastness, color gamut and colorimetric
characteristics of the final printed samples were studied for a Canon Laser Printer LBP-5050. It was found out that, the
reproducibility of the printer is statistically dependent on the paper's type. The average of color change via exposing
light is between 1.3 and 1.8 for different papers. The color difference caused by paper variation illustrates the effect of
paper on the final printed colors. The texture properties of the papers were quantified with Edge Frequency technique as
an appropriate texture analysis tool. A defined correlation between texture changes and color differences was obtained.
Furthermore, the whiteness of the papers clearly affected the final printed colors.
Printing and Halftoning II
Color-dependent banding characterization and simulation on natural document images
Show abstract
Print defects like banding from a digital press involve not only luminance variation, but also chrominance
variation. As digital presses place one color separation at a time, the contrast and spatial pattern of the print
defects are color-space dependent. Characterizing the color-dependent features of the banding signal enables us to
simulate the banding on natural document images in a more accurate way that matches the characteristics of the
banding generation mechanism within the digital press. A framework is described for color-dependent banding
characterization including the following steps: printing and scanning uniform patches that sample colorant
combinations throughout the input document sRGB color space, extracting banding signals in the CMYK color
space of the target device, and modeling the banding features in a perceptually uniform color space. We obtain
a full banding features LUT for every color point in the input sRGB space by interpolating banding features
extracted from measured color points. The color-dependent banding simulation framework is developed based
on the banding features LUT. Using the information contained in this LUT, a single banding prototype signal is
modulated in a color-space-dependent fashion that varies spatially across the natural document image. Proper
execution of the framework of banding characterization and simulation requires careful calibration of each system
component, as well as implementation of a complete color management pipeline.
Modeling large-area influence in digital halftoning for electrophotographic printers
Show abstract
Digital halftoning provides a mechanism for rendering continuous-tone images on devices such as printers. With
electrophotography, the deposition of toner within the area of a given printer addressable pixel is strongly
influenced by the halftone values of the immediately neighboring pixels. To account for these effects, it is
necessary to embed a printer model in the halftoning algorithm.
In our previous work, we used an efficient strategy to account for the impact of a 5x5 neighborhood of pixels
on the central pixel absorptance. Now we examine the potential influence of a much larger neighborhood (45x45)
of the digital halftone image on the measured value of a printed pixel at the center of that neighborhood. The
experiment shows that the extended model yields a significant improvement in the accuracy of the prediction of
the pixel values of the printed and measured halftone image.
The lattice-based screen set: a square N-color all-orders Moire-free screen set
Show abstract
Periodic clustered-dot screens are widely used for electrophotographic printers due to their print stability. However,
moir´e is a ubiquitous problem that arises in color printing due to the beating together of the clustered-dot,
periodic halftone patterns that are used to represent different colorants. This beating or interference phenomenon
introduces spurious low frequency (large period) patterns in the printed output that are very objectionable to
the viewer. The traditional solution in the graphic arts and printing industry is to rotate identical square screens
to angles that are maximally separated from each other. For example, the classic three-color screen set rotates
three identical square screens to the angles 15°, 45°, and 75°, respectively. However, the effectiveness of this
approach is limited when printing with more than four colorants, i.e. N-color printing, where N >4. Moreover,
accurately achieving the angles that have maximum angular separation requires a very high resolution plate
writer, as is used in commercial offset printing.
In this paper, we propose a systematic way to design color screen sets for periodic, clustered-dot screens
that offers more explicit control of the moir´e properties of the resulting screens when used in color printing. We
find a general concept for moir´e-free screen design that is called lattice-based screen design. The basic concept
behind our approach is the creation of the screen set on a 2-dimensional lattice in the frequency domain and
then picking each fundamental frequency vector of the individual colorant planes in the created spectral lattice
according to the desired properties. The halftone geometry of a screen set is the set of angles and frequencies
in units of lines per inch (LPI) of each screen plane. The lattice-based screen design offers more flexibility in
designing N-color screen sets with different halftone geometries, and all of them are guaranteed to be all-orders
moir´e-free. For example, by creating a square lattice in the frequency domain, square N-color moir´e-free screen
sets that consist of N rotated square screens can be achieved. The proposed approach maintains the advantage
of square clustered-dot screen design and is based on low addressability of digital printing. We also propose
several symmetry measures, and use them to compare the proposed 4-color square screen set and the screen
sets based on a previous moir´e-free N-color non-orthogonal approach. The proposed screen set is shown to have
better symmetry properties.
Colour print workflow and methods for multilayering of colour and decorative inks using UV inkjet for fine art printing
Show abstract
In order to increase density of colour and improve ink coverage when printing onto a range of non standard substrates,
this paper will present research into multi-layering of colour and the appearance of colour at 'n' levels of ink coverage.
Returning to our original investigation of artist's requirements when making inkjet prints, these observations are based
on empirical approaches that address the need to present physical data that is more useful and meaningful to the designer.
The study has used multi-pass printed colour charts to measure colour and to provide users with an understanding at a
soft-preview level to demonstrate the appearance of printed colour on different substrates. Test results relating to the
appearance of print on different surfaces, and a series of case studies will be presented using recent research into the
capabilities of UV printing technology, which has widened the opportunities for the designer to print onto non-standard
materials. It will also present a study into layering of greys and gloss in order to improve the appearance of printed
images onto metal.
Halftone blending between smooth and detail screens to improve print quality with electrophotographic printers
Show abstract
In this paper, we consider a dual-mode halftoning process for the electrophotographic laser printer - a low
frequency halftoning for smooth regions and a high frequency halftoning for detail regions. These regions are
described by an object map that is extracted from the page description language (PDL) version of the document.
This manner of switching screens depending on the local content provides a stable halftone without artifacts
in smooth areas and preserves detail rendering in detail or texture areas. However, when switching between
halftones with two different frequencies, jaggies may occur along the boundaries between areas halftoned with
low and high frequency screens. To reduce the jaggies, our screens obey a harmonic relationship. In addition, we
implement a blending process based on a transition region. We propose a nonlinear blending process in which
at each pixel, we choose the maximum of the two weighted halftones where the weights vary according to the
position in the transition region. Moreover, we describe an on-line tone-mapping for the boundary blending
process, based on an off-line calibration procedure that effectively assures the desired tone values within the
transition region.
Ink-saving strategy based on document content characterization and halftone textures
Show abstract
Common ink-saving techniques usually restrict the ink consumption when printing a document by replacing
a percentage of cyan, magenta, and yellow, by black ink. Even though such methods achieve a considerable
reduction in the amount of ink used in a page, the visual quality of the print is affected and unpleasing effects
in pastels and skin tones are observed. On the other hand, the quality of the print is not only affected by
the ink-saving algorithm, but also by the way the color halftoning algorithm arranges the dots in the print.
Therefore, the relationship between the contents of the document to be printed and the printing process needs to
be addressed by the ink-saving strategy. A color direct binary search halftoning method that strives to minimize
both the ink usage and the perceived error between the continuous-tone color image and the color halftone image
is proposed. Our goals are to estimate the effects of the ink-saving module of a printing workflow in individual
regions of the document, and to determine the dot arrangement and ink combination that consumes the least
amount of ink while preserving printing quality.
Spectral and Display
Spectral transmittance model for stacks of transparencies printed with halftone colors
Show abstract
The present work investigates the transmission of light through stacks of halftone printed transparencies. We propose a
spectral transmittance model describing the multiple reflections of light between the transparencies, whose individual
reflectance and transmittance have themselves been obtained by a prediction model. The model for single printed
transparency involves the multiple reflections of light between the interfaces as well as the orientation-dependent
attenuations of light within the plastic and ink layers. A procedure enables converting the nominal ink surface coverages
into effective ones by taking into account the spreading of the inks. Calibration of the model is based on printing a small
number of color patches on one transparency and measuring their spectral transmittance. Regarding the stacks of
transparencies, an experimental test carried out with inkjet printed samples shows good agreement between predictions
and measurements for stacks of two, three and four transparencies. Stochastic halftones are used in order to avoid the
apparition of moiré patterns when superposing the halftones. By inversion of the model, we are able to determine the
halftone colors to print on each transparency in order to obtain by superposition one targeted color. An original
application of this, called "color matching", consists in producing one color of stack from various combinations of colors
on the transparencies. The prediction accuracy of the proposed model guarantees the good visual uniformity of the
resulting colored area.
Optimal estimation of spectral reflectance based on metamerism
Show abstract
In this paper, we proposed an accurate estimation method for spectral reflectance of objects captured in an image. The
spectral reflectance is simply modeled by a linear combination of three basic spectrums of R, G, and B colors
respectively, named as spectral reflective bases of objects, which are acquired by solving a linear system based on the
principle of color metamerism. Some experiments were performed to evaluate the accuracy of the estimated spectral
reflectance of objects. The average mean square error of 24 colors in Macbeth checker between we simulated and the
measured is 0.0866, and the maximum is 0.310. In addition, the average color difference of the 24 colors is less than 1.5
under the D65 illuminant. There are 13 colors having their color difference values less than 1, and other 8 colors having
the values during the range of 1 and 2. Only three colors are relatively larger, with the differences of 2.558, 4.130 and
2.569, from the colors of No. 2, No. 13, and No. 18 in Macbeth checker respectively. Furthermore, the computational
cost of this spectral estimation is very low and suitable for many practical applications in real time.
Hue-shift model for DLP projector with the white peaking function
Show abstract
Projectors have become common display devices, not only for office and school presentations, but also for home theater
entertainment. Although a completely dark room is the ideal venue for watching a projected image, in most situations
(including classrooms and conference rooms) the viewing conditions are not completely dark, and ambient light falling
on the screen produces a background light level with the image projected on top. As the background light increases, it
becomes more difficult to see the projected image, which becomes dull and may appear washed out. What is really
happening is that the ambient light reduces the contrast of the image. While the amount of light contributing to the image
remains the same, more light has been projected onto the screen by other light sources. This effect can be reduced by
employing the white-peaking function of a digital light-processing (DLP) projector, which adjusts the white segment of
the color wheel, resulting in more natural and vivid images. Although the chromaticity coordinates for an image
projected with and without white peaking are the same, when white is added to the projected image, the perceived hue
changes. This phenomenon is known as the Abney effect. This paper presents a model of this hue-shift phenomenon and
proposes a hue-correction method. For evaluation purposes, an observer-preference test is conducted on several test
images with and without hue shifts, and z-scores are utilized to compare the results.
Content-dependent block noise reduction for mobile displays
Show abstract
Number of pixels on mobile displays is rapidly increasing. Recently, mobile displays with more than one million pixels
have been introduced into markets. However, most of multimedia contents to be displayed on mobile displays have much
smaller pixel counts. For example, number of pixels for a T-DMB(terrestrial digital multimedia broadcasting) sequence
is 320x240. When enlargement is applied to input sequence, perceived image quality would be degraded. Increase in
visibility of block noise is one of the major reasons for image quality degradation on mobile displays. This paper
presents a simple and computationally efficient method to reduce visibility of block noise on enlarged multimedia
sequences. In proposed method, a simple low pass filtering is selectively applied to the pixels of block noises for
reduction of block noise visibility as well as faithful reproduction of image details.
Poster Session
Characterization of color scanners based on SVR
Show abstract
By researching the principle of colorimetric characterization method and Support
Vector Regression (SVR), we analyze the feasibility of nonlinear transformation from scanner RGB
color space to CIELAB color space based on SVR and built a new characterization model. Then we use
the MATLABR2009a software to make a data simulation experiment to verify the accuracy of this
model and figure out the color differences by CIEDE2000 color difference formula. Based on
CIEDE2000 color difference formula, the average, the maximum and the minimum color differences of
the training set are 1.2376, 2.5593 and 0.2182, the average, the maximum and the minimum color
differences of the text set are 1.9318, 4.1421 and 0.4228. From the experimental results, we can make a
conclusion that SVR can realize the nonlinear transformation from scanner RGB color space to
CIELAB color space and the model satisfies the accuracy of scanner characterization. Therefore, SVR
can be used into the color scanner characterization management.
Deducing ink thickness variations of fluorescent print by a spectral prediction model
Show abstract
In the color printing process, the thickness and uniformity of ink have a great affect
on the color reproduction. The ink thickness uniformity is an important parameters of measuring
the quality of printing. Based on the fluorescent additives may absorb ultraviolet light and exit
blue light or visible light and by considering the expansion of the ink, optical properties of paper
with fluorescent additives , the internal lateral spread of light in paper with fluorescent additives
and the fluorescent Clapper-Yule spectral reflectance prediction model, we introduce two factor
parameters which are the initial thickness of the inks and the factor of ink thickness variation. A
model for deducing ink thickness variations of printing on the fluorescent substrate is developed
by the least square method and the spectrum reflectance of prints which measures the ink
thickness variations. The correctness of the conclusions are verified by experiment.
Bio-inspired color sketch for eco-friendly printing
Show abstract
Saving of toner/ink consumption is an important task in modern printing devices. It has a positive ecological and social
impact. We propose technique for converting print-job pictures to a recognizable and pleasant color sketches. Drawing a
"pencil sketch" from a photo relates to a special area in image processing and computer graphics - non-photorealistic
rendering. We describe a new approach for automatic sketch generation which allows to create well-recognizable
sketches and to preserve partly colors of the initial picture. Our sketches contain significantly less color dots then initial
images and this helps to save toner/ink. Our bio-inspired approach is based on sophisticated edge detection technique for
a mask creation and multiplication of source image with increased contrast by this mask. To construct the mask we use
DoG edge detection, which is a result of blending of initial image with its blurred copy through the alpha-channel, which
is created from Saliency Map according to Pre-attentive Human Vision model. Measurement of percentage of saved
toner and user study proves effectiveness of proposed technique for toner saving in eco-friendly printing mode.
Reflectance model for recto-verso color halftone images
Show abstract
In the color reproduction process, accurately predicting the color of recto-verso images and
establishing a spectral reflectance model for halftones images are the great concern project of imaging quality
control field. The scattering of light within paper and the ink penetration in the substrate are the key factors, which
affect the color reproduction. A reflectance model for recto-verso color halftone prints is introduced in this paper
which considers these factors. The paper based on the assumption that the colorant is non-scattering and the
assumption that the paper is strong scattering substrate. By the multiple internal reflection between the paper
substrate and the print-air interface of light, and the light along oblique path of the Williams-Clapper model, we
proposed the color spectral reflectance precise prediction model of recto-verso halftone images. In the study, we
propose this model for taking into account ink spreading, a phenomenon that occurs when printing an ink halftone
in superposition with one or several solid inks. The ink-spreading model includes nominal-to-effective dot area
coverage functions for each of the different ink overprint conditions by the least square curve fitting method, so
the functions for physical dot gain of various overprint halftones are given. This model provided a theoretical
foundation for color prediction analysis of recto-verso halftone images and the development of image quality
detection system.
The study on physical dot gain of second-order FM halftone based on ink spreading in all ink superposition conditions
Show abstract
Based on the feature of the second-order FM halftone dots and the interaction feature among ink, paper and air, a spectral
prediction model for the second-order FM halftone prints for the spectral Neugebauer random mixing model and
dot-on-dot mixing model is developed by using a weighting factor to signify the correlation proportion of two models.
The new spectral prediction model also includes the Yule-Nielsen effect by taking into account ink spreading in all ink
superposition conditions. We create an ink spreading function for each superposition condition to indicate the difference
of ink spreading for ink dots printing on different superposition, which maps nominal to effective dot surface coverage.
Our results shown that in the ink jet printing, the mean and max ΔE difference between measured and predicted values
are 2.90 and 7.26, smaller than the difference predicted by the model that is not taken into account ink spreading, which
are 5.52 and 12.81. The prediction accuracy is considerably improved by the new model.
Tensor decomposition for color printer model lookup table
Show abstract
Multidimensional lookup tables (LUTs) are often used to describe the response of physical systems to multiple
inputs. However these tables are also tensors, and in this paper we will use tensor decomposition to greatly reduce
the number of parameters needed to generate an accurate approximation to the tensor, and discuss how to determine
these parameters from a small number of known tensor elements. We will use this approach to generate printer
models, which are CMY or CMYK to L*a*b* LUTs where each element is an L*a*b* value for one CMYK
formulation. The approach generates accurate results with a reasonable number of L*a*b* measurements, and can
be used when nothing else is known about the system. It also runs much faster than the physics based models that
are sometimes available for these systems.
Colony image acquisition and genetic segmentation algorithm and colony analyses
W. X. Wang
Show abstract
Colony anaysis is used in a large number of engineerings such as food, dairy, beverages, hygiene, environmental
monitoring, water, toxicology, sterility testing. In order to reduce laboring and increase analysis acuracy, many
researchers and developers have made efforts for image analysis systems. The main problems in the systems are image
acquisition, image segmentation and image analysis. In this paper, to acquire colony images with good quality, an
illumination box was constructed. In the box, the distances between lights and dishe, camra lens and lights, and camera
lens and dishe are adjusted optimally. In image segmentation, It is based on a genetic approach that allow one to consider
the segmentation problem as a global optimization,. After image pre-processing and image segmentation, the colony
analyses are perfomed. The colony image analysis consists of (1) basic colony parameter measurements; (2) colony size
analysis; (3) colony shape analysis; and (4) colony surface measurements. All the above visual colony parameters can be
selected and combined together, used to make a new engineeing parameters. The colony analysis can be applied into
different applications.
Edge detection by using edge density and eleven algorithm comparisons in three types of color images
W. X. Wang,
J. Y. Xu
Show abstract
Edge detection in grey scale image processing is a traditional research subject, but recently more and more researchers
make efforts on the edge detection in color images. This paper presents a novel edge detection algorithm using the local,
nonparametric estimation of the color image density. The method firstly analyses the edge shape information provided by
the local probability distribution of the color image both in the horizontal and vertical directions respectively, then it
obtains the modulus for the edge detection in the color image. With the increasing of window size, the other types of
distributions can be simplified to the three types of the distributions presented in this study. In experiements, eleven
different edge detection algorithms are compared for the three types of color images: smooth surface objects with a few
edges; thin (or lines and curves) objects with many edges; and rough surface objects with more edges. And the
algorithms include fractional, the first and the second order differential operators and other non-differential ones.
Experiments show that the studied method is efficient.for edge extracting in a color image, and can give a satisfactory
edge detection result in most cases.
Spectral prediction model for variable dot-size ink jet presswoke
Show abstract
When ink jet printers' drop size change, based on Clapper-Yule model, study reflectance prediction model for
variable dot-size ink jet presswoke. At first, this article introduce variable drop size and the FM screening form of
Clapper-Yule model; then, using the change of ink film thickness and FM screening dot, which are caused by drop change,
extended Clapper-Yule model; at last, established the Clapper-Yule model which is extended by drop size, in addition,
validated the advantage of the model by simulation.