introduction to medical image analysis - compute.dtu.dk - week12.pdf · 21 dtu compute, technical...
TRANSCRIPT
DTU Compute
Introduction to Medical Image AnalysisRasmus R. PaulsenDTU Compute
http://courses.compute.dtu.dk/02511http://courses.compute.dtu.dk/02512
DTU Compute
Introduction to Medical Image Analysis2 DTU Compute, Technical University of Denmark
Lecture 12 – Color Images
9.00 Lecture
Spleen Challenge Finals
Course evaluation
12.00 – 13.00 Lunch break13.00 Exercises
DTU Compute
Introduction to Medical Image Analysis3 DTU Compute, Technical University of Denmark
Today's Learning Objectives Describe the basic human visual system including
rods and cones Describe subtractive colors Describe additive colors Describe the RGB color space Describe the use of the Bayer pattern in digital
cameras Convert from an RGB value to an HSI value Describe the use of different color spaces
DTU Compute
Introduction to Medical Image Analysis4 DTU Compute, Technical University of Denmark
The Human Eye
https://askabiologist.asu.edu/rods-and-cones
DTU Compute
Introduction to Medical Image Analysis5 DTU Compute, Technical University of Denmark
Color sensitivity
https://askabiologist.asu.edu/rods-and-cones
DTU Compute
Introduction to Medical Image Analysis6 DTU Compute, Technical University of Denmark
Object colorsSubtractive colors
All other colors than red absorbed
DTU Compute
Introduction to Medical Image Analysis7 DTU Compute, Technical University of Denmark
Object colorsAdditive colors
Additive colours: Final colour is made by mixing red, green, and blue
RGB = Red, Green, and Blue Television, computers, digital
cameras use the “RGB colorspace”
Typically the values of R, G, and B lie between 0 and 255
DTU Compute
Introduction to Medical Image Analysis8 DTU Compute, Technical University of Denmark
RGB Colours When alle three “Lamps” are turned of we get
black When all three “lamps” are on what do we get?
RGB = (0,0,0)
RGB = (255,255,255)
(255,0,0) (0,255,0) (0,0,255)
(255,255,0) (0,255,255)
(255,0,255)
DTU Compute
Introduction to Medical Image Analysis9 DTU Compute, Technical University of Denmark
Color camera with one sensor
http://www.skyandtelescope.com/astronomy-resources/astrophotography-tips/redeeming-color-planetary-cameras/
DTU Compute
Introduction to Medical Image Analysis10 DTU Compute, Technical University of Denmark
RGB color space
DTU Compute
Introduction to Medical Image Analysis11 DTU Compute, Technical University of Denmark
Processing RGB images Each pixel in a colour image contains 3 values Equal to a “vector function” in mathematics Much more complicated to analyse Medical images are typically grayscale Therefore we convert from colours to grayscale
before the analysis
DTU Compute
Introduction to Medical Image Analysis12 DTU Compute, Technical University of Denmark
Converting colour to grayscale
v = 0.2989 * R + 0.5870 * G + 0.1140 * B
DTU Compute
Introduction to Medical Image Analysis13 DTU Compute, Technical University of Denmark
Normalised RGB colors
DTU Compute
Introduction to Medical Image Analysis14 DTU Compute, Technical University of Denmark
Another RGB representation
(r,g,I)
DTU Compute
Introduction to Medical Image Analysis15 DTU Compute, Technical University of Denmark
HSI Color Reprentation Hue – the dominant wave length in the perceived
light (the pure color) Saturation – the purity of the color Intensity – the brightness of the color (sometimes
called the value)
DTU Compute
Introduction to Medical Image Analysis16 DTU Compute, Technical University of Denmark
Converting between RGB and HSI You have an RGB value You want the corresponding HSI value
DTU Compute
Introduction to Medical Image Analysis17 DTU Compute, Technical University of Denmark
Why other colorspaces Why should we use for example HSI ?
DTU Compute
Introduction to Medical Image Analysis18 DTU Compute, Technical University of Denmark
Melanoma segmentation An algorithm that can
do pixelwiseclassification– Background / skin– Melanoma
Use the colors
DTU Compute
Introduction to Medical Image Analysis19 DTU Compute, Technical University of Denmark
Melanoma segmentation – color variation
DTU Compute
Introduction to Medical Image Analysis20 DTU Compute, Technical University of Denmark
Color thresholding
DTU Compute
Introduction to Medical Image Analysis21 DTU Compute, Technical University of Denmark
Color thresholding
DTU Compute
Introduction to Medical Image Analysis22 DTU Compute, Technical University of Denmark
Color variation The major variation is in the brightness
– This will spread out the values in RGB space The Hue is rather constant HSI Space
– HUE and saturation rather stable– Only variation in intensity / value
DTU Compute
Introduction to Medical Image Analysis23 DTU Compute, Technical University of Denmark
Next week 3D medical image analysis Course evaluation follow up Exam questions
DTU Compute
Introduction to Medical Image Analysis24 DTU Compute, Technical University of Denmark
Exercises Segmentation of the iris
– Huge topic in biometry