DOCSLIB.ORG

RPI LRC Capturing the Lighting Edge New Color Metrics Mark Fairchild

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text

Load more

Color Appearance of Displays, etc. RGC scheduled September 28, 2012 from 6:45 AM to 9:45 AM RPI LRC Capturing the Lighting Edge New Color Metrics Oct. 3, 2012 Mark Fairchild Rochester Institute of Technology, College of Science 1 Some Adaptation Demos 2 Color 3 4 5 6 Blur (Sharpness) 7 8 9 10 Noise 11 12 13 14 Chromatic, Blur, and Noise Adaptation 15 Outline •Color Appearance Phenomena •Chromatic Adaptation •Metamerism •Color Appearance Models •HDR 16 Color Appearance Phenomena 17 Color Appearance Phenomena If two stimuli do not match in color appearance when (XYZ)1 = (XYZ)2, then some aspect of the viewing conditions differs. Various color-appearance phenomena describe relationships between changes in viewing conditions and changes in appearance. Bezold-Brücke Hue Shift Abney Effect Helmholtz-Kohlrausch Effect Hunt Effect Simultaneous Contrast Crispening Helson-Judd Effect Stevens Effect Bartleson-Breneman Equations Chromatic Adaptation Color Constancy Memory Color Object Recognition 18 Simultaneous Contrast The background in which a stimulus is presented influences the apparent color of the stimulus. Stimulus Indicates lateral interactions and adaptation. Stimulus Color- Background Background Change Appearance Change Darker Lighter Lighter Darker Red Green Green Red Yellow Blue Blue Yellow 19 Simultaneous Contrast Example (a) (b) 20 Josef Albers 21 Complex Spatial Interactions 22 Hunt Effect Corresponding chromaticities across indicated relative changes in luminance (Hypothetical Data) For a constant chromaticity, perceived 0.6 colorfulness increases with luminance. 0.5 As luminance increases, stimuli of lower colorimetric purity are required to match 1 10 a given reference stimulus. y 0.4 100 1000 10000 10000 1000 100 10 1 0.3 Indicates nonlinearities in visual processing. 0.2 0.2 0.3 0.4 0.5 0.6 x 23 Stevens Effect 1 Perceived lightness contrast increases with increasing adapting luminance. As adapting luminance increases: dark colors look darker and light colors look lighter. L (Dark) L (63 dB) Indicates luminance-dependent nonlinearities. L (79 dB) L (97 dB) 0.1 0.1 1 Y/Yn 24 Stevens & Hunt Effects 0.1 cd/m2 1.0 cd/m2 10 cd/m2 100 cd/m2 1000 cd/m2 10,000 cd/m2 25 Bartleson-Breneman Perceived lightness as a function of relative luminance for various surround relative luminances 100 90 Apparent contrast in complex stimuli (i.e. images) 80 increases with increasing surround luminance. 70 60 Decreased surround luminance increases the 50 brightness of all image colors, but the effect is 40 greater for dark colors. 30 Lightness (Average) 20 Lightness (Dim) Indicates a differential contrast effect (white-point 10 Lightness (Dark) resetting). 0 0 10 20 30 40 50 60 70 80 90 100 Relative Luminance, Y 26 Surround Effect Demo 27 28 Adaptation Light Adaptation: Decrease in visual sensitivity with increases in luminance. (Automatic Exposure Control) Dark Adaptation: Increase in visual sensitivity with decreases in luminance. (Automatic Exposure Control) Chromatic Adaptation: Independent sensitivity regulation of the mechanisms of color vision. (Automatic Color Balance) 29 Local Adaptation Linear Mapping Perceptual Mapping 30 Chromatic Adaptation 1.25 The three cone types, LMS, are capable of 1 independent sensitivity regulation. (Adaptation occurs in higher-level mechanisms as well.) 0.75 Magnitudes of chromatic responses are dependent 0.5 on the state of adaptation (local, spatial, temporal). Afterimages provide evidence. 0.25 0 400 450 500 550 600 650 700 Wavelength (nm) 31 32 A Cyan Filter 33 34 35 Color Constancy (Discounting) 0.54 Adapting Chromaticities Achromatic Chromaticities Incandescent We perceive the colors of objects to remain 0.52 unchanged across large changes in illumination <-- Hands <-- No Hands color. •Not True v' 0.50 •Chromatic Adaptation •Poor Color Memory •Cognitive Discounting-the-Illuminant 0.48 Daylight <-- Hands <-- No Hands 0.46 0.18 0.20 0.22 0.24 0.26 u' 36 Diamonds 37 With Tips 38 On Backgrounds 39 Both Tips & Backgrounds! 40 Real-World Discounting 41 42 Chromatic Adaptation Modeling Chromatic Adaptation: Largely independent sensitivity regulation of the (three) mechanisms of color vision. 1.25 1 0.75 0.5 0.25 0 400 450 500 550 600 650 700 Wavelength (nm) 43 Chromatic Adaptation Models Model: X1Y1Z1 La = f(L, Lwhite ,...) 3x3 Ma = f(M, Mwhite,...) L1M1S1 S a = f(S, Swhite,...) VC1 LaMaSa Transform (CAT): VC2 L M S XYZ 2 = f(XYZ1, XYZwhite ,...) 2 2 2 3x3 X2Y2Z2 44 XYZ-to-LMS 2 z 1 S M L 1.5 y 0.75 s t i e v i u l t i a s L 0.400 0.708 −0.081 X v n x e y s s u 0.5 1 l e u v i m t i t a l M = −0.226 1.165 0.046 Y s i e r R 0.25 T S 0.000 0.000 0.918 Z 0.5 0 380 480 580 680 780 Wavelength, nm 0 380 480 580 680 780 Wavelength, nm 45 Johannes von Kries Johannes von Kries "Father of Chromatic-Adaptation Models" 46 von Kries Hypothesis "This can be conceived in the sense that the individual components present in the organ of vision are completely independent of one another and each is fatigued or adapted exclusively according to its own function." -von Kries, 1902 La = kLL Ma = k MM S = k S a S Von Kries thought of this “proportionality law” as an extension of Grassmann’s Laws to span two viewing conditions. 47 Some Evolution of CATs •Nayatani et al. Nonlinear Model •Fairchild (1991) & (1994) Models •Bradford Model •CIELAB & CIELUV •CAT02 48 Metamerism: Illuminant, Observer, & Metrics 49 Illuminant Metamerism • Same Observer, Change Illuminantion • In theory, must have tristimulus equality for a reference illuminant. • Calculate Color Difference for Test Illuminant: CIE Special Index of Metamerism • Calculate Spectral Difference: General Index of Metamerism 50 In Practice, Paramers, not Metamers Corrected Standard Batch batch X 19.93 21.96 19.93 Y 14.29 14.87 14.29 Z 7.23 7.91 7.23 0.8 HJ HJ HJ J 0.6 H r o t Standard c B a f e J J Batch c n 0.4 H a t B H Corrected batch c e B l f B e B R B J B 0.2 H B HJ HJ H H J B B BJ H J J BHJ H H BHJ BHJ BJ B B 0 400 450 500 550 600 650 700 Wavelength (nm) 51 CIE Special Index of Metamerism, Change in Illuminant • Must have an exact tristimulus match for a single viewing and observer condition. • Calculate Color Difference for Test Illuminant • Workflow: • Measure Parameric Pair • Correct Batch as Needed for Reference Illuminant • Calculate CIE DE2000 for Test Illuminant 52 General Index of Metamerism is Spectral Based • Weighted RMS Error (Using Diagonal of Matrix R) n ⎛ ⎞2 ⎜ w ΔR ⎟ ∑⎝ λ λ ⎠ WRMS = λ=1 n w = diag R λ ( ) 53 Observer Metamerism 1931 2° vs. 1964 10° standard observers 2 1.5 s e u l a v s u 1 l u m i t s i r T 0.5 0 380 480 580 680 780 Wavelength, nm Dashed lines, 1964 observer 54 Variability in Color Matching Functions Wyszecki and Stiles, “Color Science” 55 CIE Standard Deviate Observer 2.2 • Based on Stiles and Burch 1.8 data s e u 1.4 l a v s Normalization of primaries u 1 l • u and V-lambda minimized m i t s i 0.6 r true variance T 0.2 Not recommended -0.2 • 380 480 580 680 780 Wavelength, nm • CIE Pub. 80 20-year old (solid lines) 60-year old (dashed lines) 56 CIE TC 1-36 (2006) ⎣⎡− Dτ ,max,macula ⋅Dmacula,relative (λ)− Dτ ,ocul (λ)⎦⎤ l (λ) = αi,l (λ)⋅10 ⎣⎡− Dτ ,max,macula ⋅Dmacula,relative (λ)− Dτ ,ocul (λ)⎦⎤ m(λ) = αi,m (λ)⋅10 ⎣⎡− Dτ ,max,macula ⋅Dmacula,relative (λ)− Dτ ,ocul (λ)⎦⎤ s (λ) = αi,s (λ)⋅10 Color-matching Cone spectral Macula spectral Lens and ocular functions absorptances density media spectral as a function of density as a field size function of age (i=internal) 57 Examples of Cone Fundamentals L-, M-, & S-Cone Fundamentals (2- & 10-Deg. @ Age 32) L-, M-, & S-Cone Fundamentals (10-Deg. @ Ages 20 & 80) 1.00 1.00 l-bar (2) l-bar (20) 0.90 m-bar (2) 0.90 m-bar (20) s-bar (2) s-bar (20) 0.80 l-bar (10) 0.80 l-bar (80) m-bar (10) m-bar (80) 0.70 0.70 s-bar (10) s-bar (80) 0.60 0.60 0.50 0.50 0.40 0.40 Relative Sensitivity Relative 0.30 Sensitivity Relative 0.30 0.20 0.20 0.10 0.10 0.00 0.00 390 440 490 540 590 640 690 740 390 440 490 540 590 640 690 740 Wavelength (nm) Wavelength (nm) Normalized to Peak Height 58 Effect of Age vs. Field Size 2 2 xbar 1931 xbar 1931 ybar 1931 ybar 1931 zbar 1931 zbar 1931 x25yrs,2degree x25,10degree 1.5 1.5 y25yrs,2degree y25,10degree z25yrs,2degree z25,10degree x65yrs,2degree x65,10degree y65yrs,2degree y65,10degree 1 1 z65yrs,2degree z65,10degree ristimulus values ristimulus values T 0.5 T 0.5 0 0 -0.5 -0.5 380 430 480 530 580 630 680 730 780 380 430 480 530 580 630 680 730 780 Wavelength (nm) Wavelength (nm) 59 Some Metamers 0.7 0.8 Vinyl-1 O Vinyl-1 G 0.7 0.6 Vinyl 2-A Vinyl 2-B 0.6 Fabric-A 0.5 Fabric-B 0.5 0.4 0.4 0.3 Reflectance factor Reflectance factor 0.3 0.2 0.2 0.1 0.1 0 0 380 430 480 530 580 630 680 730 780 380 430 480 530 580 630 680 730 780 Wavelength (nm) Wavelength (nm) Grays Materials 60 Observer Metamerism 2,25 2,65 1931 10,25 61 10,65 Illuminant Metamerism 1931, D65 1931, A 1931, HP Na 1964, HP Na 62 Color “Blindness” 63 Cone Spectral Sensitivities 1 S M L y 0.75 t i v i t i s n e s 0.5 e v i t a l e R 0.25 0 380 480 580 680 780 Wavelength, nm 64 Color Deficiencies Dichromats Protanopia: Lacking “L-cone response” Deuteranopia: Lacking “M-cone response” Tritanopia: Lacking “S-cone response” Also “Anomalous Trichromats” ~8% of Males have Color Vision Deficiencies Various tests available at: <www.richmondproducts.com>, <www.munsell.com>.
Recommended publications
  • When Red Lights Look Yellow

    When Red Lights Look Yellow

    Publications 11-2005 When Red Lights Look Yellow Joanne M. Wood Queensland University of Technology David A. Atchison Queensland University of Technology Alex Chaparro Wichita State University, [email protected] Follow this and additional works at: https://commons.erau.edu/publication Part of the Cognition and Perception Commons, Musculoskeletal, Neural, and Ocular Physiology Commons, and the Ophthalmology Commons Scholarly Commons Citation Wood, J. M., Atchison, D. A., & Chaparro, A. (2005). When Red Lights Look Yellow. Investigative Ophthalmology & Visual Science, 46(11). https://doi.org/10.1167/iovs.04-1513 This Article is brought to you for free and open access by Scholarly Commons. It has been accepted for inclusion in Publications by an authorized administrator of Scholarly Commons. For more information, please contact [email protected]. When Red Lights Look Yellow Joanne M. Wood,1 David A. Atchison,1 and Alex Chaparro2 PURPOSE. Red signals are typically used to signify danger. This observer’s distance spectacle prescription. It did not occur study was conducted to investigate a situation identified by with lens powers in excess of ϩ1.00 D, as the signals became train drivers in which red signals appear yellow when viewed too blurred for the viewer to distinguish the color. A compre- at long distances (ϳ900 m) through progressive-addition hensive eye and vision examination of the train driver who had lenses. originally reported the color misperception revealed that his METHODS. A laboratory study was conducted to investigate the corrected vision was normal. The train driver was also shown effects of defocus, target size, ambient illumination, and sur- to have normal color vision, as assessed by the Ishihara, Farns- round characteristics on the extent of the color misperception worth Lantern, and Farnsworth D15 tests.
  • Urban Land Grab Or Fair Urbanization?

    Urban Land Grab Or Fair Urbanization?

    Urban land grab or fair urbanization? Compulsory land acquisition and sustainable livelihoods in Hue, Vietnam Stedelijke landroof of eerlijke verstedelijking? Landonteigenlng en duurzaam levensonderhoud in Hue, Vietnam (met een samenvatting in het Nederlands) Chiếm đoạt đất đai đô thị hay đô thị hoá công bằng? Thu hồi đất đai cưỡng chế và sinh kế bền vững ở Huế, Việt Nam (với một phần tóm tắt bằng tiếng Việt) Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op maandag 21 december 2015 des middags te 12.45 uur door Nguyen Quang Phuc geboren op 10 december 1980 te Thua Thien Hue, Vietnam Promotor: Prof. dr. E.B. Zoomers Copromotor: Dr. A.C.M. van Westen This thesis was accomplished with financial support from Vietnam International Education Development (VIED), Ministry of Education and Training, and LANDac programme (the IS Academy on Land Governance for Equitable and Sustainable Development). ISBN 978-94-6301-026-9 Uitgeverij Eburon Postbus 2867 2601 CW Delft Tel.: 015-2131484 [email protected]/ www.eburon.nl Cover design and pictures: Nguyen Quang Phuc Cartography and design figures: Nguyen Quang Phuc © 2015 Nguyen Quang Phuc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission in writing from the proprietor. © 2015 Nguyen Quang Phuc.
  • Color Difference Delta E - a Survey

    Color Difference Delta E - a Survey

    See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/236023905 Color difference Delta E - A survey Article in Machine Graphics and Vision · April 2011 CITATIONS READS 12 8,785 2 authors: Wojciech Mokrzycki Maciej Tatol Cardinal Stefan Wyszynski University in Warsaw University of Warmia and Mazury in Olsztyn 157 PUBLICATIONS 177 CITATIONS 5 PUBLICATIONS 27 CITATIONS SEE PROFILE SEE PROFILE All content following this page was uploaded by Wojciech Mokrzycki on 08 August 2017. The user has requested enhancement of the downloaded file. Colour difference ∆E - A survey Mokrzycki W.S., Tatol M. {mokrzycki,mtatol}@matman.uwm.edu.pl Faculty of Mathematics and Informatics University of Warmia and Mazury, Sloneczna 54, Olsztyn, Poland Preprint submitted to Machine Graphic & Vision, 08:10:2012 1 Contents 1. Introduction 4 2. The concept of color difference and its tolerance 4 2.1. Determinants of color perception . 4 2.2. Difference in color and tolerance for color of product . 5 3. An early period in ∆E formalization 6 3.1. JND units and the ∆EDN formula . 6 3.2. Judd NBS units, Judd ∆EJ and Judd-Hunter ∆ENBS formulas . 6 3.3. Adams chromatic valence color space and the ∆EA formula . 6 3.4. MacAdam ellipses and the ∆EFMCII formula . 8 4. The ANLab model and ∆E formulas 10 4.1. The ANLab model . 10 4.2. The ∆EAN formula . 10 4.3. McLaren ∆EMcL and McDonald ∆EJPC79 formulas . 10 4.4. The Hunter color system and the ∆EH formula . 11 5. ∆E formulas in uniform color spaces 11 5.1.
  • Several Color Appearance Phenomena in Color Reproduction

    Several Color Appearance Phenomena in Color Reproduction

    2nd International Conference on Electronic & Mechanical Engineering and Information Technology (EMEIT-2012) Several Color Appearance Phenomena in Color Reproduction Qin-ling Dai1,a, Xiao-zhou Li2*,b, Ai Xu2 1 School of Materials Engineering (Southwest Forestry University), Kunming, China 650224 2 Key Laboratory of Pulp & Paper Science and Technology (Shandong Polytechnic University), Ministry of Education, Ji’nan, China, 250353 [email protected], bcorresponding author: [email protected] Keywords: color reproduction, color appearance, color appearance phenomena Abstract. Color perceived performance was influenced by various color appearance phenomena caused by varying viewing conditions in color reproduction process. It is necessary to do some research on the color appearance phenomena to represent the color appearance models qualitatively and quantitatively and accurate color reproduction easily. Only the phenomena were studied thoroughly, could the color transmission and reproduction be well performed. The color appearance and common color appearance phenomena of color reproduction were analyzed in this paper. And the basic theory of color appearance in color reproduction was also studied. Introduction High fidelity digital printing plays an important role in high fidelity color transmission and reproduction and it is one of the most important techniques to perform high fidelity color reproduction. High fidelity digital printing helps to perform accurate color reproduction of the original which can’t be performed because of paper and ink in traditional printing process [1]. In color printing, the color difference caused by paper, ink and viewing conditions is various. The difference is not only colorimetric difference but also different color appearance phenomena. While the different color appearance phenomenon is the leading factor to influence the color vision perceived.
  • The Art of Digital Black & White by Jeff Schewe There's Just Something

    The Art of Digital Black & White by Jeff Schewe There's Just Something

    The Art of Digital Black & White By Jeff Schewe There’s just something magical about watching an image develop on a piece of photo paper in the developer tray…to see the paper go from being just a blank white piece of paper to becoming a photograph is what many photographers think of when they think of Black & White photography. That process of watching the image develop is what got me hooked on photography over 30 years ago and Black & White is where my heart really lives even though I’ve done more color work professionally. I used to have the brown stains on my fingers like any good darkroom tech, but commercially, I turned toward color photography. Later, when going digital, I basically gave up being able to ever achieve what used to be commonplace from the darkroom–until just recently. At about the same time Kodak announced it was going to stop making Black & White photo paper, Epson announced their new line of digital ink jet printers and a new ink, Ultrachrome K3 (3 Blacks- hence the K3), that has given me hope of returning to darkroom quality prints but with a digital printer instead of working in a smelly darkroom environment. Combine the new printers with the power of digital image processing in Adobe Photoshop and the capabilities of recent digital cameras and I think you’ll see a strong trend towards photographers going digital to get the best Black & White prints possible. Making the optimal Black & White print digitally is not simply a click of the shutter and push button printing.
  • Accurately Reproducing Pantone Colors on Digital Presses

    Accurately Reproducing Pantone Colors on Digital Presses

    Accurately Reproducing Pantone Colors on Digital Presses By Anne Howard Graphic Communication Department College of Liberal Arts California Polytechnic State University June 2012 Abstract Anne Howard Graphic Communication Department, June 2012 Advisor: Dr. Xiaoying Rong The purpose of this study was to find out how accurately digital presses reproduce Pantone spot colors. The Pantone Matching System is a printing industry standard for spot colors. Because digital printing is becoming more popular, this study was intended to help designers decide on whether they should print Pantone colors on digital presses and expect to see similar colors on paper as they do on a computer monitor. This study investigated how a Xerox DocuColor 2060, Ricoh Pro C900s, and a Konica Minolta bizhub Press C8000 with default settings could print 45 Pantone colors from the Uncoated Solid color book with only the use of cyan, magenta, yellow and black toner. After creating a profile with a GRACoL target sheet, the 45 colors were printed again, measured and compared to the original Pantone Swatch book. Results from this study showed that the profile helped correct the DocuColor color output, however, the Konica Minolta and Ricoh color outputs generally produced the same as they did without the profile. The Konica Minolta and Ricoh have much newer versions of the EFI Fiery RIPs than the DocuColor so they are more likely to interpret Pantone colors the same way as when a profile is used. If printers are using newer presses, they should expect to see consistent color output of Pantone colors with or without profiles when using default settings.
  • Multispectral Image Reproduction Via Color Appearance Mapping

    Multispectral Image Reproduction Via Color Appearance Mapping

    International Journal of Signal Processing, Image Processing and Pattern Recognition Vol.7, No.4 (2014), pp.65-72 http://dx.doi.org/10.14257/ijsip.2014.7.4.06 Multispectral Image Reproduction via Color Appearance Mapping Ying Wang, Sheping Zhai and Zhongmin Wang School of Computer Science & Technology, Xi’an Univ. of Posts & Telecommunications, Xi’an China [email protected] Abstract To achieve the color consistent reproduction of multispectral images in the different viewing condition, a new method of multispectral image reproduction via color appearance mapping was proposed. Firstly, through the introduction of color appearance transformation, the color appearance description of the source spectral reflectance in the source viewing condition was obtained. Then by the construction of the inverse model, the high dimension spectra in the destination viewing condition were evaluated, which was color appearance matching but spectral mismatching with the source image. Finally, to improve the spectral precision of the reproduced spectra, the evaluated spectra were corrected by the method of metamerism correction based on the source spectra, and then the reproduced spectral image was obtained, which matched the source image in color appearance and in spectra when the reproduction viewing condition was different from the source. Experiments show that the perceptual color difference and the spectral error between the reproduced multi-spectral image and its original in the different viewing condition are small. The new method preserves the spectral information of the source multispectral image and achieves equal perceptual reproduction to the source image. Keywords: multispectral image reproduction, color appearance mapping, spectral evaluation, metamerism correction, viewing condition independent space 1.
  • Predictability of Spot Color Overprints

    Predictability of Spot Color Overprints

    Predictability of Spot Color Overprints Robert Chung, Michael Riordan, and Sri Prakhya Rochester Institute of Technology School of Print Media 69 Lomb Memorial Drive, Rochester, NY 14623, USA emails: [email protected], [email protected], [email protected] Keywords spot color, overprint, color management, portability, predictability Abstract Pre-media software packages, e.g., Adobe Illustrator, do amazing things. They give designers endless choices of how line, area, color, and transparency can interact with one another while providing the display that simulates printed results. Most prepress practitioners are thrilled with pre-media software when working with process colors. This research encountered a color management gap in pre-media software’s ability to predict spot color overprint accurately between display and print. In order to understand the problem, this paper (1) describes the concepts of color portability and color predictability in the context of color management, (2) describes an experimental set-up whereby display and print are viewed under bright viewing surround, (3) conducts display-to-print comparison of process color patches, (4) conducts display-to-print comparison of spot color solids, and, finally, (5) conducts display-to-print comparison of spot color overprints. In doing so, this research points out why the display-to-print match works for process colors, and fails for spot color overprints. Like Genie out of the bottle, there is no turning back nor quick fix to reconcile the problem with predictability of spot color overprints in pre-media software for some time to come. 1. Introduction Color portability is a key concept in ICC color management.
  • Fast and Stable Color Balancing for Images and Augmented Reality

    Fast and Stable Color Balancing for Images and Augmented Reality

    Fast and Stable Color Balancing for Images and Augmented Reality Thomas Oskam 1,2 Alexander Hornung 1 Robert W. Sumner 1 Markus Gross 1,2 1 Disney Research Zurich 2 ETH Zurich Abstract This paper addresses the problem of globally balanc- ing colors between images. The input to our algorithm is a sparse set of desired color correspondences between a source and a target image. The global color space trans- formation problem is then solved by computing a smooth Source Image Target Image Color Balanced vector field in CIE Lab color space that maps the gamut of the source to that of the target. We employ normalized ra- dial basis functions for which we compute optimized shape parameters based on the input images, allowing for more faithful and flexible color matching compared to existing RBF-, regression- or histogram-based techniques. Further- more, we show how the basic per-image matching can be Rendered Objects efficiently and robustly extended to the temporal domain us- Tracked Colors balancing Augmented Image ing RANSAC-based correspondence classification. Besides Figure 1. Two applications of our color balancing algorithm. Top: interactive color balancing for images, these properties ren- an underexposed image is balanced using only three user selected der our method extremely useful for automatic, consistent correspondences to a target image. Bottom: our extension for embedding of synthetic graphics in video, as required by temporally stable color balancing enables seamless compositing applications such as augmented reality. in augmented reality applications by using known colors in the scene as constraints. 1. Introduction even for different scenes. With today’s tools this process re- quires considerable, cost-intensive manual efforts.
  • Colornet--Estimating Colorfulness in Natural Images

    Colornet--Estimating Colorfulness in Natural Images

    COLORNET - ESTIMATING COLORFULNESS IN NATURAL IMAGES Emin Zerman∗, Aakanksha Rana∗, Aljosa Smolic V-SENSE, School of Computer Science, Trinity College Dublin, Dublin, Ireland ABSTRACT learning-based objective metric ‘ColorNet’ for the estimation of colorfulness in natural images. Based on a convolutional neural Measuring the colorfulness of a natural or virtual scene is critical network (CNN), our proposed ColorNet is a two-stage color rating for many applications in image processing field ranging from captur- model, where at stage I, a feature network extracts the characteristics ing to display. In this paper, we propose the first deep learning-based features from the natural images and at stage II, a rating network colorfulness estimation metric. For this purpose, we develop a color estimates the colorfulness rating. To design our feature network, rating model which simultaneously learns to extracts the pertinent we explore the designs of the popular high-level CNN based fea- characteristic color features and the mapping from feature space to ture models such as VGG [22], ResNet [23], and MobileNet [24] the ideal colorfulness scores for a variety of natural colored images. architectures which we finally alter and tune for our colorfulness Additionally, we propose to overcome the lack of adequate annotated metric problem at hand. We also propose a rating network which dataset problem by combining/aligning two publicly available color- is simultaneously learned to estimate the relationship between the fulness databases using the results of a new subjective test which characteristic features and ideal colorfulness scores. employs a common subset of both databases. Using the obtained In this paper, we additionally overcome the challenge of the subjectively annotated dataset with 180 colored images, we finally absence of a well-annotated dataset for training and validating Col- demonstrate the efficacy of our proposed model over the traditional orNet model in a supervised manner.
  • Method for Compensating for Color Differences Between Different Images of a Same Scene

    Method for Compensating for Color Differences Between Different Images of a Same Scene

    (19) TZZ¥ZZ__T (11) EP 3 001 668 A1 (12) EUROPEAN PATENT APPLICATION (43) Date of publication: (51) Int Cl.: 30.03.2016 Bulletin 2016/13 H04N 1/60 (2006.01) (21) Application number: 14306471.5 (22) Date of filing: 24.09.2014 (84) Designated Contracting States: (72) Inventors: AL AT BE BG CH CY CZ DE DK EE ES FI FR GB • Sheikh Faridul, Hasan GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO 35576 CESSON-SÉVIGNÉ (FR) PL PT RO RS SE SI SK SM TR • Stauder, Jurgen Designated Extension States: 35576 CESSON-SÉVIGNÉ (FR) BA ME • Serré, Catherine 35576 CESSON-SÉVIGNÉ (FR) (71) Applicants: • Tremeau, Alain • Thomson Licensing 42000 SAINT-ETIENNE (FR) 92130 Issy-les-Moulineaux (FR) • CENTRE NATIONAL DE (74) Representative: Browaeys, Jean-Philippe LA RECHERCHE SCIENTIFIQUE -CNRS- Technicolor 75794 Paris Cedex 16 (FR) 1, rue Jeanne d’Arc • Université Jean Monnet de Saint-Etienne 92443 Issy-les-Moulineaux (FR) 42023 Saint-Etienne Cedex 2 (FR) (54) Method for compensating for color differences between different images of a same scene (57) The method comprises the steps of: - for each combination of a first and second illuminants, applying its corresponding chromatic adaptation matrix to the colors of a first image to compensate such as to obtain chromatic adapted colors forming a chromatic adapted image and calculating the difference between the colors of a second image and the chromatic adapted colors of this chromatic adapted image, - retaining the combination of first and second illuminants for which the corresponding calculated difference is the smallest, - compensating said color differences by applying the chromatic adaptation matrix corresponding to said re- tained combination to the colors of said first image.
  • The Helmholtz-Kohlrausch Effect

    The Helmholtz-Kohlrausch Effect

    Out of the Wood BY MIKE WOOD Lightness— The Helmholtz-Kohlrausch effect Don’t be put off by the strange name also strongly suggest looking at the online lightness most people see when looking at of this issue’s article. The Helmholtz- version of this article, as the images will be this image. Kohlrausch (HK) effect might sound displayed on your monitor and behave Note: Not everyone will see the differences esoteric, but it’s a human eye behavioral more like lights than the ink pigments in as strongly, particularly with a printed page effect with colored lighting that you are the printed copy. You can access Protocol like this. Just about everyone sees this effect undoubtedly already familiar with, if issues on-line at http://plasa.me/protocol with lights, but the strength of the effect varies perhaps not under its formal name. This or through the iPhone or iPad App at from individual to individual. In particular, effect refers to the human eye (or entoptic) http://plasa.me/protocolapp. if you are red-green color-blind, then you may phenomenon that colored light appears The simplest way to show the Helmholtz- see very little difference. brighter to us than white light of the same Kohlrausch effect is with an illustration. What I see—which will agree with what luminance. This is particularly relevant to The top half of Figure 1 shows seven most of you see—is that the red and pink the entertainment industry, as it is most differently colored patches against a grey patches look by far the brightest, while blue, obvious when using colored lights.