Multimedia Systems - Colour Video � Colour is a visual Presence and distributions � of colours induce feature which is sensations and conveys immediately perceived meanings in the observer according to specific rules Joemon Jose Representing colour on � Salient chromatic � digital images and properties are www.dcs.gla.ac.uk/~ jj/teaching/demms4/ reproducing accurately on captured output devices are not at � Colour can add great all straightforward value to an image Tuesday, 15 th January 2008 Distances in colour space � should correspond to human perceptual distance DEMS4 2004 DEMS4 2004 Video Video 2 14/01/04 1 14/01/04 Human Information Human Visual Perception processing � Identical colour combinations can cause different � Mixing three primary colours in varying colour sensation under different conditions proportions, the perception of different � Likewise two different colour can be perceived identical … colours can be created � the human eye & brain � Human eye build up of � Interpolation � Cones to perceive colour � Pictures and events that can still be identified as separate � By exciting retina using different intensities of � Colour interaction in the brain the three primary colours, the same colour may � Adaptation be perceived by the brain even if its unique � General-brightness adaptation wavelength is not present. Lateral adaptation � Chromatic adaptation � DEMS4 2004 DEMS4 2004 Video Video 3 4 14/01/04 14/01/04
Colour Space CIE Chromaticity Diagram � To deal with colour we need to quantify it in some way � gives us the notion of colour space or domain � Hierarchy of colour sets � Perceivable by human beings � Displayed on a monitor screen � Calculated and stored in a frame memory DEMS4 2004 DEMS4 2004 Video Video 14/01/04 5 14/01/04 6 Video Technology: Representation of Colour representing colour Stimuli � Points in three � Hardware-oriented monochrome � dimensional space models � bilevel � one bit/pixel: 0 = black, 1 = white � Calorimetric models � RGB, CMY, YIQ � grey-scale � User-oriented models � CIE Chromaticity � e.g., 8 bits/pixel = 256 intensities diagram � HLS, HSV, HSB colour � Physiologically inspired � � value for each colour gun models � no of bits gives colour range � CIE XYZ, RGB � e.g., 24 bits = 8 bits for red, 8 bits for green, 8 bits � Psychological models for blue � HSV, � colour depth DEMS4 2004 DEMS4 2004 Video Video 7 8 14/01/04 14/01/04
Video Technology: Video Technology: generating a colour Colour Models: RGB frame buffer � RGB = Red Green Blue (2 D array of � directly modelled in device (i.e., corresponds to phosphor dots 24 bit values) colour guns on display colour guns in display) � easy to implement red � not based on visual (perceived) colours <128, 128, 255> green � not perceptually uniform What RGB value you 8 bits per colour blue see DEMS4 2004 DEMS4 2004 Video Video 14/01/04 9 14/01/04 10 Video Technology: Video Technology: Colour Models: RGB Colour Space Colour Models: RGB Colour Space -,-,z Cyan Blue Blue (0,1,1) (0,0,1) Cyan -,y,- X,-,- Magenta Magenta White White (1,0,1) (1,1,1) Black Black Green (0,0,0) Green (0,1,0) Red Yellow Yellow Red (1,0,0) (1,1,0) DEMS4 2004 DEMS4 2004 Video Video 11 12 14/01/04 14/01/04
Video Technology: Video Technology: Colour Models: RGB Colour Models: HSV � Colour is labeled as a relative weights of three � HSV = hue, saturation, value (intensity) primary colours, in an additive system using the � “painter’s model” primaries Red, Green, Blue � better model for representing colours as we see � It is perceptually non-linear space them (“I want a bright highly saturated apple � Equal distances in the space do not necessarily green.”) correspond to perceptually equal sensation � desaturation = adding white � Non-linear relationship between RGB values & the � can be converted to/from RGB intensity produced in each phosphor dot, low � like RGB, axes not perceptually uniform intensity values produce small changes in response to screen � variant: HLS (hue, lightness, saturation) � It is not a good colour description system DEMS4 2004 DEMS4 2004 Video Video 14/01/04 13 14/01/04 14 Video Technology: Video Technology: Colour Models: HSV Colour Space Colour Models: HSV Non-linear transformation of RGB cube � Hue : quality by which we distinguish one family from others � V Chroma: quality by which we distinguish a strong colour from � Green Yellow weak ones Value: It is that quality by which we distinguish a light colour � from a dark one Cyan Red H corresponds to selecting a colour; S corresponds to � selecting the amount of white; selecting V corresponds to Blue adding black Magenta Perceptually non-linear � Perceptual in the sense that we are using attributes that we � normally think of h Attributes are not independent s � variant: HLS (hue, lightness, saturation) � DEMS4 2004 DEMS4 2004 Video Video 15 16 14/01/04 14/01/04
Video Technology: Video Technology: Colour Models: YUV CIE Colour Specification System � colour model used for TV signal transmission � Commission Internationale d’Éclairage � Y represents luminance (intensity of monochrome � colour labelling system signal) � “XYZ” space � U,V carry separate colour information (colour � international standard (1931) difference values) � based on colour matching functions determined by � Y = 0.2125R + 0.7154G + 0.0721B experiments with human subjects � U = B-Y, V = R-Y � gives uniform colour spaces � typically, Y contributes most to signal bandwidth � needs transformation into one of the other models DEMS4 2004 DEMS4 2004 Video Video 14/01/04 17 14/01/04 18 Video Technology: Pushing the hardware Colour Models: CMYK � CMYK = cyan, magenta, yellow, black � Consumers expectations are based on broadcast television � “printer’s model” � Consumer equipment plays back at reduced � a subtractive model frame rate resulting in jittery- dropped � set of practically available CMYK colours frames (“process colours”) are not equivalent to � In order to accommodate low-end PCs RGB set considerable compromises over quality must be made DEMS4 2004 DEMS4 2004 Video Video 19 20 14/01/04 14/01/04
Human Perception Persistence of vision � If a sequence of still images is presented to our � What frame rate perceived as smooth? eyes at sufficiently high rate (frame rate~40 fps), � No identification of single frames if refresh we experience a continuous visual sensation rather frequency is high enough than perceiving individual images � Perception of 16 frames/s as continuous sequence � A lag in the eye’s response to visual stimuli which results � Depends on material in after images � More sensitive to low frequencies � If the consecutive images only differ by a small amount, any changes from one to next will be � More sensitive to changes in luminance perceived as movement of elements within images and blue-orange axis � Film projector displays an image twice (24 fps � Vision emphasizes edge detection becomes 48 fps) DEMS4 2004 DEMS4 2004 Video Video 14/01/04 21 14/01/04 22 Digitization: camera vs Video Sequence computer Consists of number of frames � � Images produced by digitising time-varying signal Advantage � � disadvantage generated by the sensors in a camera � Analogue signal � Bit-mapped images � User has no control transmitted on a Camera over digitization cable get corrupted � by noise � Most conform to an � Circuitry Inside a Camera appropriate standard � Noise will creep in if � Purely digital signal (data stream) is fed into a computer analogue data is via a high speed interface stored on a magnetic � IEEE 1394 (FireWire) tape Computer � � Camera is resistant � Broadcast video is fed into a video capture card attached to corruption by to the computer noise and � Video capture card- analogue signal is converted into a interference digital form DEMS4 2004 DEMS4 2004 Video Video 23 24 14/01/04 14/01/04
Recommend
More recommend
