Held: Monday, 23 September 2013

Back to Outline 13 - Pause for Breath. On to Outline 15 - Transforming Images.

Summary

Today we explore common ways in which to transform colors, pixels, and images.

Related Pages

• Reading: Transforming RGB Colors
• Lab: Transforming RGB Colors
• EBoard (Source) (PDF)

Overview

• Review: Color basics.
• Computing new colors from old.
• An example.

Administrivia

• Study break tonight at 8 pm.
• My other class wants the "Current EBoard" link to give the source file, rather than the HTML. Do you care?
• Any parents coming Friday? (I'm trying to figure out how to fit them.)
• I've gotten a few "I don't think I should ask you this" emails. If you are asking questions via email, I don't consider anything out of bounds. (I may refuse to answer, but I won't be bothered that you've asked.)
• New lab partners!
• Lots of folks had trouble with the GIMP tools stuff. I didn't need exact code, but I wanted to be able to have a sense that you understood the "choose color" -> "select area" -> "fill" order of things.
• I'm going to have an optional class this Thursday at 1:15 (and, I hope, most Thursdays) as a chance to talk to folks about more things that might be useful or interesting.
• For those who missed the documentation lab writeup, here's an alternative. Document the following procedure

(define range
  (lambda (val cap)
    (+ 1 (modulo (- val 1) cap))))

• Work for Tuesday
  • Continue to work on the exam
  • Reading for Tuesday: Transforming Images
  • Lab writeup 4: Problems 6a and 7b from Transforming RGB Colors.
    Due before class on Wednesday. (Ideally before class on Tuesday.)
• EC Opportunities:
  • CS Extras Thursday @ 4:30: Jennelle Nystrom on working at Microsoft
  • CS Table Friday (The Story of Mel)
  • Other? (Water polo, Football, G-Tones, ...)

Review: RGB Colors

• A standard way to represent transmissive colors.
• Think of it as a combination of three colored lights of varying intensities
  • Red
  • Green
  • Blue
• Why these colors? It just works that way.
• How do we measure intensities?
  
  • Some people like a 0-1 scale
  • Computers like a 0-255 scale (powers of two rock!)
• Important procedures:
  • (rgb-new red green blue) - create a new color
  • (image-set-pixel! image col row color) - set the color of a pixel
  • (image-get-pixel image col row) - get the color of a pixela
  • (rgb-red color) - get the red component
  • (rgb-green color) - get the green component
  • (rgb-blue color) - get the blue component
  • (color->rgb-list color) - get a list of the three components
  • (color->rgb color) - convert some representations to the RGB representation.

Transforming Colors

• Today we are exploring ways to transform colors.
• Why?
  • Deepens our understanding of colors.
  • We can apply a color transformation to a pixel in the image.
  • We can apply a color transformation to each pixel in an image. (More on that tomorrow.)

An Example

• MediaScript comes with a variety of built-in color transformations.
• However, we can also define our own (yay!).
• The normal framework for a color transformation is

(define color-transform
  (lambda (color)
    (rgb-new ___ ; computation of red component
             ___ ; computation of green component
             ___ ; computation of blue component
             )))

• Let's consider a simple one: We'll keep only the blue component of an image.
• How would we write this?

(define only-blue
  (lambda (color)
    (rgb-new ___
             ___
             ___)))

• Some tests
• While we'll spend today using this idea with pixels, for this example, we'll explore the effect on a sample image.