DOCSLIB.ORG

Color Management Fundamentals Wide Format Series

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Color Management Fundamentals Wide Format Series Color Management Fundamentals Wide Format Series Kerry Moloney John Nate Field & Channel Marketing Manager – Fiery Wide Format WW Technical Product Training Manager – Fiery Wide Format Session overview • Color management fundamentals, including basic color theory and management terms • Get the best out of all of color technologies, regardless of the technology type • Focus today on printing, with an emphasis on some wide format printing considerations • Take a quick poll 2 Topics to be covered • Color management defined • Color space • Color accuracy and Delta-E (ΔΕ or dE) • Device dependent and device independent color • Gamut • Rendering intents & black point compensation • Color transforms • Calibration & profiling • Spot colors 3 What is color management? And why do we need it? 4 Color management is… • A process, not a task • Ensures color fidelity between devices • Limited by each device 5 Color management is… • A process, not a task • Ensures color fidelity between devices • Limited by each device 6 Color management is… • A process, not a task • Ensures color fidelity between devices • Limited by each device 7 Color management is… …like making toast… by Steve Upton 8 Color management is… …like making toast… by Steve Upton 9 Color management is… …like making toast… 10 Color management is… …like making toast… 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 11 Color management fundamentals Basic terms 12 Color space (L*a*b*) 100 50 L* 0 13 Color space (L*a*b*) -127 0 +127 a* 14 Color space (L*a*b*) -128 0 +127 b* 15 Color space (L*a*b*) L* b* a* 16 Color space (L*a*b*) L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 17 Black Color space (L*a*b*) L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 18 Black Color inaccuracy • Hue - color • Chroma - saturation • Luminance - brightness 19 Hue Cooler Actual Warmer Color 20 Where is your color? L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 21 Black Hue L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 22 Black Hue L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 23 Black Chroma - Saturation Less Actual More Saturation Color Saturation 24 Chroma or Saturation L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 25 Black Chroma or Saturation L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 26 Black Chroma or Saturation L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 27 Black Luminance Darker Actual Lighter Color 28 Where is your color? L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 29 Black Where is your color? L* = 100 White b* Yellow -a* a* Gray Green Red -b* Blue L* = 0 30 Black Delta-E or ΔΕ or De Original Color Delta-E 3 Delta-E 7 31 Device dependent color 32 Device dependent color R = 201 R = 206 G = 5 G = 3 B = 2 B = 7 33 Device dependent color CMYK = 1/84/95/2 5/93/88/0 0/97/90/8 34 Device dependent color CMYK = 1/84/95/2 5/93/88/0 0/97/90/8 L*a*b* = 56/95/54 35 Gamut 36 Gamut 37 Gamut 38 Gamut 39 Gamut 40 Gamut 41 Gamut 42 Gamut 43 Color space SWOP 44 Color space SWOP vs. GRACoL 45 Color space SWOP vs. GRACoL vs. Inkjet 46 Color space SWOP vs. GRACoL vs. Inkjet vs. Adobe RGB 47 Color space SWOP vs. Toner vs. Inkjet 48 Gamut Larger Smaller 49 Gamut ? Larger Smaller 50 Spot colors 51 Spot colors – Pantone Bridge 52 Color management application Decision time 53 Rendering intent Larger Gamut Smaller Gamut In-Gamut Colors 4 5 6 1 3 2 Out-Of-Gamut Colors 54 Perceptual 1 3 4 5 6 2 55 Perceptual 1 3 4 5 6 2 56 Perceptual 1 3 4 5 6 2 57 Perceptual 1 3 4 5 6 2 58 Perceptual 1 3 4 5 6 2 59 Perceptual 1 3 4 5 6 2 60 Perceptual 1 3 4 5 6 2 61 Colorimetric 1 3 4 5 6 2 62 Saturation 1 3 4 5 6 2 63 Colorimetric – relative vs. absolute Newsprint Absolute Colorimetric Newsprint Relative Colorimetric 64 Banding Original Perceptual Colorimetric 65 Banding Original Perceptual Colorimetric 66 Banding Original Perceptual Colorimetric 67 Banding Original Perceptual Colorimetric 68 Black point compensation Newsprint 69 Black point compensation 70 Black point compensation 71 Black point compensation 72 Graphic program defaults 73 Graphic program defaults 74 Assigning vs. converting 75 Assigning vs. converting Assigning = Change the way the image looks/prints Converting = Keep the look the same 76 Assigning vs. converting Assign = Not the same taste 1 Tb. = 1 Tb. 77 Assigning vs. converting Convert = The same taste 1 Tb. = 4 Tb. 78 Assigning vs. converting Original Color in GRACoL 23/92/53/17 40/50/14 79 Assigning vs. converting Original Color in GRACoL 23/92/53/17 40/50/14 Assign SWOP 5 23/92/53/17 41/51/10 80 Assigning vs. converting Original Color in GRACoL 23/92/53/17 40/50/14 Assign SWOP 5 Convert SWOP 5 23/92/53/17 24/93/61/19 41/51/10 40/50/14 81 Converting GRACoL SWOP 82 Assigning vs. converting Assign Assign Convert 83 Calibration and profiling Controlling your color 84 Calibration • Calibration is active 85 Calibration • Calibration is active • Match a device to itself over time 86 Calibration • Calibration is active • Match a device to itself over time • Bring to a repeatable state 87 Calibration • Calibration is active • Match a device to itself over time • Bring to a repeatable state • Tuning the guitar 88 Calibration • Sometimes called “linearization” • Linearization is only one part of calibration 89 Calibration – cameras • Shutter speed • Aperture • White/gray balance 90 Calibration – monitors • Brightness • Contrast • White point 91 Calibration – output devices 92 Calibration – output devices • Per channel ink/toner limits • Per channel linearity • Total ink/toner limit • Ink gray balance • Usually through the use of a RIP 93 Calibration – output devices • Per channel ink limits 94 Calibration – output devices • Per channel ink limits Wet No additional saturation No additional saturation Wet 95 Calibration – output devices • Per channel ink limits Wet 92% No additional saturation 89% No additional saturation 96% Wet 81% 96 Calibration – output devices • Per channel linearity (ink or toner) 75% 97 Calibration – output devices • Per channel linearity (ink or toner) 75% 25% 98 Calibration – output devices • Per channel linearity (ink or toner) Non-Linear 75% 25% Non-Linear 99 Calibration – output devices • Per channel linearity (ink or toner) Non-Linear 75% 25% Non-Linear 25% 50% 75% Linear 100 Calibration – output devices • Total ink limit No darker 280% 290% 300% 310% 320% 330% 340% 350% 360% 370% 380% 390% 400% 101 Calibration – output devices • Total ink limit No darker 280% 290% 300% 310% 320% 330% 340% 350% 360% 370% 380% 390% 400% TIL 102 Calibration – output devices • Total ink limit No darker 280% 290% 300% 310% 320% 330% 340% 350% 360% 370% 380% 390% 400% TIL Wet 280% 290% 300% 310% 320% 330% 340% 350% 360% 370% 380% 390% 400% 103 Calibration – output devices • Total ink limit No darker 280% 290% 300% 310% 320% 330% 340% 350% 360% 370% 380% 390% 400% TIL Wet 280% 290% 300% 310% 320% 330% 340% 350% 360% 370% 380% 390% 400% TIL 104 Calibration – output devices • Ink gray balance Poor gray balance Questionable Gray Balance Good Gray Balance 105 Calibration – output devices • Ink gray balance Poor gray balance Questionable gray balance Good Gray Balance 106 Calibration – output devices • Ink gray balance Poor gray balance Questionable gray balance Good gray balance 107 Profiling • Profiling is passive ICC Profile International Color Consortium 108 Profiling • Profiling is passive • A snapshot of the device characteristics 109 Profiling CMYK 0/100/66/0 0/51/24/0 66/36/24/0 92/0/81/23 0/0/0/30 L*a*b* 61/127/105 73/36/13 55/-10/-18 46/-55/16 78/0/0 110 Profiling CMYK = 1/84/95/2 0/100/66/0 0/97/90/8 CMYK = 0/100/66/0 L*a*b* = 61/127/105 L*a*b* = 61/127/105 111 Color management is… …like making toast… 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 112 Session summary • What we learned – Basic color management concepts apply to getting the most out of all types of color technologies – Take charge of your color management setting and options – Color management is not a single task, rather a complete process • To learn more, we have some resources for you 113 Additional resources • www.efi.com • ABCs of Color • Rendering Intent • Black Point Compensation articles • www.colorwiki.com • Glossary of color terms • www.xrite.com • Complete Guide to Color Management • A Guide to Understanding Color Communication • The Color Guide and Glossary 114 Additional World of Fiery sessions • May 27 – Color Management Fundamentals • Learn the foundations of color theory and management to get the most from your design software and color technologies. • June 17 – Color Standards & Specifications • Gain knowledge of the latest industry standards and how they apply to wide and superwide format printing and proofing. • July 8 – Color Management Inside Fiery XF & Fiery proServer • Put your knowledge into practice and understand RIP-level color controls. • Production Series for wide format printing starting this September, you can register your interest for these on the World of Fiery Webinars home page 115 Color management fundamentals Questions? 116 Color management fundamentals Thank you! 117.
Load more

Recommended publications
  • Color Management
    Color Management hotoshop 5.0 was justifiably praised as a ground- breaking upgrade when it was released in the summer of 1998, although the changes made to the color P management setup were less well received in some quarters. This was because the revised system was perceived to be complex and unnecessary. Bruce Fraser once said of the Photoshop 5.0 color management system ‘it’s push-button simple, as long as you know which of the 60 or so buttons to push!’ Attitudes have changed since then (as has the interface) and it is fair to say that most people working today in the pre-press industry are now using ICC color profile managed workflows. The aim of this chapter is to introduce the basic concepts of color management before looking at the color management interface in Photoshop and the various color management settings. 1 Color management Adobe Photoshop CS6 for Photographers: www.photoshopforphotographers.com The need for color management An advertising agency art buyer was once invited to address a meeting of photographers. The chair, Mike Laye, suggested we could ask him anything we wanted, except ‘Would you like to see my book?’ And if he had already seen your book, we couldn’t ask him why he hadn’t called it back in again. And if he had called it in again we were not allowed to ask why we didn’t get the job. And finally, if we did get the job we were absolutely forbidden to ask why the color in the printed ad looked nothing like the original photograph! That in a nutshell is a problem which has bugged many of us throughout our working lives, and it is one which will be familiar to anyone who has ever experienced the difficulty of matching colors on a computer display with the original or a printed output.
    [Show full text]
  • Image Processing
    IMAGE PROCESSING ROBOTICS CLUB SUMMER CAMP’12 WHAT IS IMAGE PROCESSING? IMAGE PROCESSING = IMAGE + PROCESSING WHAT IS IMAGE? IMAGE = Made up of PIXELS Each Pixels is like an array of Numbers. Numbers determine colour of Pixel. TYPES OF IMAGES 1. BINARY IMAGE 2. GREYSCALE IMAGE 3. COLOURED IMAGE BINARY IMAGE Each Pixel has either 1 (White) or 0 (Black) Depth =1 (bit) Number of Channels = 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 1 1 1 1 1 0 0 0 0 1 1 1 1 1 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 GRAYSCALE Each Pixel has a value from 0 to 255. 0 : black and 1 : White Between 0 and 255 are shades of b&w. Depth=8 (bits) Number of Channels =1 GRAYSCALE IMAGE RGB IMAGE Each Pixel stores 3 values :- R : 0- 255 G: 0 -255 B : 0-255 Depth=8 (bits) Number of Channels = 3 RGB IMAGE HSV IMAGE Each pixel stores 3 values :- H ( hue ) : 0 -180 S (saturation) : 0-255 V (value) : 0-255 Depth = 8 (bits) Number of Channels = 3 Note : Hue in general is from 0-360 , but as hue is 8 bits in OpenCV , it is shrinked to 180 STARTING WITH OPENCV OpenCV is a library for C language developed for Image Processing To embed opencv library in Dev C complier , follow instructions in :- http://opencv.willowgarage.com/wiki/DevCpp HEADER FILES IN C After embedding openCV library in Dev C include following header files:- #include "cv.h" #include "highgui.h" IMAGE POINTER An image is stored as a structure IplImage with following elements :- int height Width int width int nChannels int depth Height char *imageData int widthStep ….
    [Show full text]
  • Preparing Images for Delivery
    TECHNICAL PAPER Preparing Images for Delivery TABLE OF CONTENTS So, you’ve done a great job for your client. You’ve created a nice image that you both 2 How to prepare RGB files for CMYK agree meets the requirements of the layout. Now what do you do? You deliver it (so you 4 Soft proofing and gamut warning can bill it!). But, in this digital age, how you prepare an image for delivery can make or 13 Final image sizing break the final reproduction. Guess who will get the blame if the image’s reproduction is less than satisfactory? Do you even need to guess? 15 Image sharpening 19 Converting to CMYK What should photographers do to ensure that their images reproduce well in print? 21 What about providing RGB files? Take some precautions and learn the lingo so you can communicate, because a lack of crystal-clear communication is at the root of most every problem on press. 24 The proof 26 Marking your territory It should be no surprise that knowing what the client needs is a requirement of pro- 27 File formats for delivery fessional photographers. But does that mean a photographer in the digital age must become a prepress expert? Kind of—if only to know exactly what to supply your clients. 32 Check list for file delivery 32 Additional resources There are two perfectly legitimate approaches to the problem of supplying digital files for reproduction. One approach is to supply RGB files, and the other is to take responsibility for supplying CMYK files. Either approach is valid, each with positives and negatives.
    [Show full text]
  • Converting an Alpha Channel to a Spot Channel
    43_589164 bk09ch01.qxd 6/2/05 11:12 AM Page 629 Chapter 1: Prepping Graphics for Print In This Chapter ߜ Picking the right resolution, mode, and format ߜ Prepress and working with a service bureau ߜ Creating color separations reparing images for the screen is a snap compared to what you have to Pgo through to get images ripe for the printing process. If all you ever want to do is print your images to a desktop laser or inkjet printer, the task is a little easier, but you still must take some guidelines into account. And prepping your images for offset printing? Well, throw in an additional set of guidelines. It’s not rocket science, mind you. If you stick to the basic rules and, more importantly, spend some time developing a good work- ing relationship with your service bureau and offset printer, you’re good to go. Getting the Right Resolution, Mode, and Format If you’re not familiar with the concept of resolution, I suggest taking a look at Book II, Chapter 1. That’s where I cover all the basics on resolution, pixel dimension, resampling, and other related topics. For full descriptions on color modes and file formats, see Book II, Chapter 2. That said, the next few sections give you the lowdownCOPYRIGHTED on the proper settings MATERIALfor an image that will ultimately go to print. Resolution and modes Table 1-1 provides some guidelines on what resolution settings to use for the most common type of output. Remember these are just guidelines. They aren’t chiseled into stone to withstand the sands of time or anything lofty like that.
    [Show full text]
  • Image Formats
    Image Formats Ioannis Rekleitis Many different file formats • JPEG/JFIF • Exif • JPEG 2000 • BMP • GIF • WebP • PNG • HDR raster formats • TIFF • HEIF • PPM, PGM, PBM, • BAT and PNM • BPG CSCE 590: Introduction to Image Processing https://en.wikipedia.org/wiki/Image_file_formats 2 Many different file formats • JPEG/JFIF (Joint Photographic Experts Group) is a lossy compression method; JPEG- compressed images are usually stored in the JFIF (JPEG File Interchange Format) >ile format. The JPEG/JFIF >ilename extension is JPG or JPEG. Nearly every digital camera can save images in the JPEG/JFIF format, which supports eight-bit grayscale images and 24-bit color images (eight bits each for red, green, and blue). JPEG applies lossy compression to images, which can result in a signi>icant reduction of the >ile size. Applications can determine the degree of compression to apply, and the amount of compression affects the visual quality of the result. When not too great, the compression does not noticeably affect or detract from the image's quality, but JPEG iles suffer generational degradation when repeatedly edited and saved. (JPEG also provides lossless image storage, but the lossless version is not widely supported.) • JPEG 2000 is a compression standard enabling both lossless and lossy storage. The compression methods used are different from the ones in standard JFIF/JPEG; they improve quality and compression ratios, but also require more computational power to process. JPEG 2000 also adds features that are missing in JPEG. It is not nearly as common as JPEG, but it is used currently in professional movie editing and distribution (some digital cinemas, for example, use JPEG 2000 for individual movie frames).
    [Show full text]
  • Basics of Color Management
    Application Notes Basics of Color Management Basics of Color Management ErgoSoft AG Moosgrabenstr. 13 CH-8595 Altnau, Switzerland © 2010 ErgoSoft AG, All rights reserved. The information contained in this manual is based on information available at the time of publication and is sub- ject to change without notice. Accuracy and completeness are not warranted or guaranteed. No part of this manual may be reproduced or transmitted in any form or by any means, including electronic me- dium or machine-readable form, without the expressed written permission of ErgoSoft AG. Brand or product names are trademarks of their respective holders. The ErgoSoft RIP is available in different editions. Therefore the description of available features in this document does not necessarily reflect the license details of your edition of the ErgoSoft RIP. For information on the features included in your edition of the ErgoSoft RIPs refer to the ErgoSoft homepage or contact your dealer. Rev. 1.1 Basics of Color Management i Contents Introduction ................................................................................................................................................................. 1 Color Spaces ................................................................................................................................................................ 1 Basics ........................................................................................................................................................................ 1 CMYK .......................................................................................................................................................................
    [Show full text]
  • Color Images, Color Spaces and Color Image Processing
    color images, color spaces and color image processing Ole-Johan Skrede 08.03.2017 INF2310 - Digital Image Processing Department of Informatics The Faculty of Mathematics and Natural Sciences University of Oslo After original slides by Fritz Albregtsen today’s lecture ∙ Color, color vision and color detection ∙ Color spaces and color models ∙ Transitions between color spaces ∙ Color image display ∙ Look up tables for colors ∙ Color image printing ∙ Pseudocolors and fake colors ∙ Color image processing ∙ Sections in Gonzales & Woods: ∙ 6.1 Color Funcdamentals ∙ 6.2 Color Models ∙ 6.3 Pseudocolor Image Processing ∙ 6.4 Basics of Full-Color Image Processing ∙ 6.5.5 Histogram Processing ∙ 6.6 Smoothing and Sharpening ∙ 6.7 Image Segmentation Based on Color 1 motivation ∙ We can differentiate between thousands of colors ∙ Colors make it easy to distinguish objects ∙ Visually ∙ And digitally ∙ We need to: ∙ Know what color space to use for different tasks ∙ Transit between color spaces ∙ Store color images rationally and compactly ∙ Know techniques for color image printing 2 the color of the light from the sun spectral exitance The light from the sun can be modeled with the spectral exitance of a black surface (the radiant exitance of a surface per unit wavelength) 2πhc2 1 M(λ) = { } : λ5 hc − exp λkT 1 where ∙ h ≈ 6:626 070 04 × 10−34 m2 kg s−1 is the Planck constant. ∙ c = 299 792 458 m s−1 is the speed of light. ∙ λ [m] is the radiation wavelength. ∙ k ≈ 1:380 648 52 × 10−23 m2 kg s−2 K−1 is the Boltzmann constant. T ∙ [K] is the surface temperature of the radiating Figure 1: Spectral exitance of a black body surface for different body.
    [Show full text]
  • Which Image Format
    CASE STUDY Which Image Format PNG PNG (Portable Network Graphics), an extensible file format for the lossless, portable, well-compressed storage of raster images. PNG provides a patent-free replacement for GIF and can also replace many common uses of TIFF. Indexed-color, gray scale, and true color images are supported, plus an optional alpha channel. Sample depths range from 1 to 16 bits. However, the format is not widely supported in common image programs. AVI AVI stands for Audio Video Interleave and is currently the most common file format for storing audio/video data on the PC. AVI’s are 8-bit per image plane. This file format conforms to the Microsoft Windows Resource Interchange File Format (RIFF) specification, which makes it convenient for sharing the image sequence between computers. AVI files (which typically end in the .avi extension) require a specific player that supports. RAW A RAW image format contains minimally processed data from the image sensor. RAW literally means unprocessed or uncooked. RAW images must be processed and converted to RGB format if it is a color image. Photron however, does not limit RAW as a unprocessed image. The “Bayer” check box must be selected to save the RAW image as an unprocessed image. RAW images have 8-bits per image plane. RAWW A RAWW image format contains minimally processed data from the image sensor. RAWW images must be processed and converted to RGB format if it is a color image. Photron however, does not limit RAWW as a unprocessed image. The “Bayer” check box must be selected to save the RAWW image as an unprocessed image.
    [Show full text]
  • Deep Learning-Based Color Holographic Microscopy
    Deep learning-based color holographic microscopy Tairan Liu1,2,3†, Zhensong Wei1†, Yair Rivenson1,2,3†,*, Kevin de Haan1,2,3, Yibo Zhang1,2,3, Yichen Wu1,2,3, and Aydogan Ozcan1,2,3,4,* † Equally contributing authors * Corresponding authors: [email protected] ; [email protected] 1Electrical and Computer Engineering Department, University of California, Los Angeles, CA, 90095, USA. 2Bioengineering Department, University of California, Los Angeles, CA, 90095, USA. 3California NanoSystems Institute (CNSI), University of California, Los Angeles, CA, 90095, USA. 4Department of Surgery, David Geffen School of Medicine, University of California, Los Angeles, CA, 90095, USA. Abstract We report a framework based on a generative adversarial network (GAN) that performs high-fidelity color image reconstruction using a single hologram of a sample that is illuminated simultaneously by light at three different wavelengths. The trained network learns to eliminate missing-phase-related arti- facts, and generates an accurate color transformation for the reconstructed image. Our framework is ex- perimentally demonstrated using lung and prostate tissue sections that are labeled with different histo- logical stains. This framework is envisaged to be applicable to point-of-care histopathology, and pre- sents a significant improvement in the throughput of coherent microscopy systems given that only a sin- gle hologram of the specimen is required for accurate color imaging. 1. INTRODUCTION Histological staining of fixed, thin tissue sections mounted on glass slides is one of the fundamental steps required for the diagnoses of various medical conditions. Histological stains are used to highlight the constituent tissue parts by enhancing the colorimetric contrast of cells and subcellular components for microscopic inspection.
    [Show full text]
  • Color Management Handbook
    Color Management Handbook Strategies to master color management in the digital workflow. Start applying them today. ver.5 Is that really the correct color? “Is this color good to go?” — A hesitation we often have before making prints in the digital workflow. Photographer Designer Are the application settings on the monitor correctly adjusted Is the image displayed on the monitor really accurate? and does the color match the printed image? Retoucher Printer Is the photograph edited the way it was intended? Do the colors in the design comp and color proof match? 2 Color Management Handbook 3 4 Color Management Handbook Management Color 5 makes it possible to handle data smoothly. data handle to possible it makes photographic, design, and plate making stages, and making it the shared standard, standard, shared the it making and stages, making plate and design, photographic, Maintaining an awareness of the final printed color in the finished product in the the in product finished the in color printed final the of awareness an Maintaining coincide, but color management can make them approximate one another. another. one approximate them make can management color but coincide, color gamuts that can be reproduced. These two gamuts cannot be made to to made be cannot gamuts two These reproduced. be can that gamuts color printing color standard of ISO coated v2, we can tell that there is a difference in the the in difference a is there that tell can we v2, coated ISO of standard color printing work step. work If we compare the color space widely
    [Show full text]
  • Arxiv:1902.00267V1 [Cs.CV] 1 Feb 2019 Fcmue Iin Oto H Aaesue O Mg Classificat Image Th for in Used Applications Datasets Fundamental the Most of the Most Vision
    ColorNet: Investigating the importance of color spaces for image classification⋆ Shreyank N Gowda1 and Chun Yuan2 1 Computer Science Department, Tsinghua University, Beijing 10084, China [email protected] 2 Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055, China [email protected] Abstract. Image classification is a fundamental application in computer vision. Recently, deeper networks and highly connected networks have shown state of the art performance for image classification tasks. Most datasets these days consist of a finite number of color images. These color images are taken as input in the form of RGB images and clas- sification is done without modifying them. We explore the importance of color spaces and show that color spaces (essentially transformations of original RGB images) can significantly affect classification accuracy. Further, we show that certain classes of images are better represented in particular color spaces and for a dataset with a highly varying number of classes such as CIFAR and Imagenet, using a model that considers multi- ple color spaces within the same model gives excellent levels of accuracy. Also, we show that such a model, where the input is preprocessed into multiple color spaces simultaneously, needs far fewer parameters to ob- tain high accuracy for classification. For example, our model with 1.75M parameters significantly outperforms DenseNet 100-12 that has 12M pa- rameters and gives results comparable to Densenet-BC-190-40 that has 25.6M parameters for classification of four competitive image classifica- tion datasets namely: CIFAR-10, CIFAR-100, SVHN and Imagenet. Our model essentially takes an RGB image as input, simultaneously converts the image into 7 different color spaces and uses these as inputs to individ- ual densenets.
    [Show full text]
  • 14. Color Mapping
    14. Color Mapping Jacobs University Visualization and Computer Graphics Lab Recall: RGB color model Jacobs University Visualization and Computer Graphics Lab Data Analytics 691 CMY color model • The CMY color model is related to the RGB color model. •Itsbasecolorsare –cyan(C) –magenta(M) –yellow(Y) • They are arranged in a 3D Cartesian coordinate system. • The scheme is subtractive. Jacobs University Visualization and Computer Graphics Lab Data Analytics 692 Subtractive color scheme • CMY color model is subtractive, i.e., adding colors makes the resulting color darker. • Application: color printers. • As it only works perfectly in theory, typically a black cartridge is added in practice (CMYK color model). Jacobs University Visualization and Computer Graphics Lab Data Analytics 693 CMY color cube • All colors c that can be generated are represented by the unit cube in the 3D Cartesian coordinate system. magenta blue red black grey white cyan yellow green Jacobs University Visualization and Computer Graphics Lab Data Analytics 694 CMY color cube Jacobs University Visualization and Computer Graphics Lab Data Analytics 695 CMY color model Jacobs University Visualization and Computer Graphics Lab Data Analytics 696 CMYK color model Jacobs University Visualization and Computer Graphics Lab Data Analytics 697 Conversion • RGB -> CMY: • CMY -> RGB: Jacobs University Visualization and Computer Graphics Lab Data Analytics 698 Conversion • CMY -> CMYK: • CMYK -> CMY: Jacobs University Visualization and Computer Graphics Lab Data Analytics 699 HSV color model • While RGB and CMY color models have their application in hardware implementations, the HSV color model is based on properties of human perception. • Its application is for human interfaces. Jacobs University Visualization and Computer Graphics Lab Data Analytics 700 HSV color model The HSV color model also consists of 3 channels: • H: When perceiving a color, we perceive the dominant wavelength.
    [Show full text]