C280 Computer VisionC280, Computer VisionProf. Trevor [email protected]@eecs.berkeley.eduLt3ClLecture 3: ColorColorColor`ColorColorReadings:Fhd P Ch 6–Forsyth and Ponce, Chapter 6– Szeliski, 2.3.2Last time• Image formation affect ed by geometry, photometry, and optics.• Projection equations express how world points mapped to 2d image.• Homogenous coordinates allow linear system for projection equations.• Lenses make pinhole model practical• Photometry models: Lambertian, BRDFy,• Digital imagers, Bayer demosaicingParameters (focal length aperture lens diameterParameters (focal length, aperture, lens diameter, sensor sampling…) strongly affect image obtained.K. GraumanSlide CreditsSlide Credits•KristenGrauman: 3‐48 50‐75 79‐86Kristen Grauman: 348, 5075, 7986• Bob Woodham: 49, 87‐90d h idi l (S l i• and others, indirectly (Steve Palmer, Brian Wandell, etc!)Today: Color• Measuring color– Spectral power distributions– Color mixing– Color matching experiments–Color spaces• Uniform color spaces• Perception of colorHuman photoreceptors–Human photoreceptors– Environmental eff ects, adaptation• Using color in machine vision systemsColor and light• Color of light arriving at camera depends on–Spectral reflectance of the surface light is leaving–Spectral reflectance of the surface light is leaving– Spectral radiance of light falling on that patch• Color perceived depends onPh i f li ht–Physics of light– Visual system receptors–Brain processing, environmentColor and lightWhite light: gcomposed of about equal energy in all q gywavelengths of the visible spectrumpNewton 1665Image from http://micro.magnet.fsu.edu/Electromagnetic spectrumImage credit: nasa.govHuman Luminance Sensitivity FunctionMeasuring spectraSpectroradiometer: separate input light into its different wavelengths, and measure the energy Foundations of Vision, B. Wandellat each.Spectral power distribution• The power per unit area at each wa velength of a radiant objectof a radiant object# Photons(per ms.)(p )400 500 600 700Wavelength (nm )Wavelength (nm.)Figure © Stephen E. Palmer, 2002Spectral power distributionsSome examples of the spectra of light sourcesB. Gallium Phosphide CrystalotonsA. Ruby Laserotons400 500 600 700# Pho400 500 600 700# PhoD. Normal DaylightWavelength (nm.)Wavelength (nm.)C. Tungsten LightbulbPhotonsPhotons# P400 500 600 700# P400 500 600 700© Stephen E. Palmer, 2002The color viewed is also affected by the surface’s tl fl t tispectral reflectance properties.Spectral reflectances for some natural bj t h hobjects: how much of each wavelength is reflected for that fsurfaceForsyth & Ponce, measurements by E. KoivistoSurface reflectance spectraSome examples of the reflectance spectra of surfaceseflectedRedYellow Blue Purpletons Re% PhotWavelength (nm)400 700 400 700 400 700 400 700© Stephen E. Palmer, 2002The Psychophysical CorrespondenceThere is no simple functional description for the perceivedThere is no simple functional description for the perceivedcolor of all lights under all viewing conditions, but …...A helpful constraint:A helpful constraint:Consider only physical spectra with normal distributionsmeanarea# Photonsvariance400 700500 600varianceWavelength (nm.)© Stephen E. Palmer, 2002The Psychophysical CorrespondenceMeanHueMeanHueonsyellowgreenbluePhoto# Wavelength© Stephen E. Palmer, 2002The Psychophysical CorrespondenceVarianceSaturationVarianceSaturationhighhi.onsmediummed.lPhotolowlow# Wavelength© Stephen E. Palmer, 2002The Psychophysical CorrespondenceAreaBrightnessAreaBrightnessBAreaLightnessonsB.AreaLightnessPhotobright# darkWavelength© Stephen E. Palmer, 2002Color mixingCartoon spectra for color names:Source: W. FreemanAdditive color mixingColors combine by yadding color spectraLight adds to black.Source: W. FreemanExamples of additive color systemsCRT phosphorsmultiple projectorshttp://www.jegsworks.comhttp://www.crtprojectors.co.uk/Superposition• Additive mixing: gThe spectral power distribution of the mixture is the sum of the spectral power di ib i f hdistributions of the components.Figure from B. Wandell, 1996Subtractive color mixingColors combine by multiplyingcolormultiplyingcolor spectra.Pigments removecolor from incident light (white).Source: W. FreemanExamples of subtractive color systems• Printing on paper• Crayons•Most photogra phic filmMost photographic filmToday: Color• Measuring color– Spectral power distributions– Color mixing– Color matching experiments–Color spaces• Uniform color spaces• Perception of colorHuman photoreceptors–Human photoreceptors– Environmental eff ects, adaptation• Using color in machine vision systemsWhy specify color numerically?• Accurate color reproduction is commercially valuable – Many products are identified by color (“golden” arches)• Few color names are widely recognized by English speakers– 11: black, blue, brown, grey, green, orange, pink, purple, red, white, and yellow.y– Other languages have fewer/more.– Common to disagree on appropriate color names.• Color reproduction problems increased by prevalence of digital imaging –e.g. digital libraries of art. –How to ensure that everyone perceives the same color?y p– What spectral radiances produce the same response from people under simple viewing conditions?Forsyth & PonceColor matching experiments• Goal: find out what spectral radiances produce same response in human observersColor matching experimentsObserver adjusts weight (intensity) for primary lights (fixed SPD’s) to match appearance of test li hFoundations of Vision, by Brian Wandell, Sinauer Assoc., 1995After Judd & Wyszecki.light.Color matching experiments• Goal: find out what spectral radiances produce same response in human observers• Assumption: simple viewing conditions, where p p g ,we say test light alone aff ects perception–Ignoring additional factors for now like adaptation,Ignoring additional factors for now like adaptation, complex surrounding scenes, etc.Color matching experiment 1Color matching experiment 1Slide credit: W. FreemanColor matching experiment 1Color matching experiment 1p1 p2 p3Slide credit: W. FreemanColor matching experiment 1Color matching experiment 1p1 p2 p3Slide credit: W. FreemanColor matching experiment 1Color matching experiment 1The primary color