Class 21: Images as Functions from Positions to Colors

Held: Monday, 27 February 2017

We consider yet another model of creating images, one in which the color of a pixel depends only on its position

Preliminaries

Overview

• Models of images
• Iterating over positions; images as functions
• Blends and other positionall-computed images
• Computing simple shapes
• Some strange computations

Related Pages

• Eboard 21 (Source)

• Reading: Building Images by Iterating Over Positions

• Lab: Building Images by Iterating Over Positions

Updates

News / Etc.

• New partners!

Rotating reminders

• Use our tutors! We have tutors available Sunday through Thursday evening from 7-10 p.m. in Science 3813/15.

Upcoming Work

• Assignment 5 due TOMORROW at 10:30 p.m.
• Lab writeup: 3e and 3f. (Don’t start with these; Doing the prior problems helps.)
• Reading: Revisiting lists
• Exam 2 arrives Wednesday. Yay!

Extra credit (Academic/Artistic)

• CS Table, Tuesday at noon: Facial Recognition “algorithms”
• W. Kamau Bell, Tuesday at 6pm, Harris.
• Prazak Quartet, Tuesday, 28 February 2017, 7:30 p.m., Herrick.
• CS Extras, Thursday at 4:15 pm, Sam’s research and PM’s research

Extra credit (Peer)

Extra credit (Misc)

• Preventing sexual assault, Tuesday at 11am, JRC 101
• Host a prospective student THIS COMING WEEKEND. (March 5, April 8, April 23)

Good things to do

Models of images

• We’ve seen a variety of models of images.
  • Images that we describe by composing basic shapes.
  • Images that we create by transforming existing images.
  • Images that we create by using the Gimp tools.
  • Images that we create by a combination of techniques.
• But there are certainly other models of image making. We’ll learn at least two more this semester.
• Why are we learning so many models of image making? Because the basic operations you have affect the way you think about problem solving.

The image-compute model

• In the image-variant model, we built images with a starting image and a function from colors to colors.
• We will consider a variant of this model - images as functions from positions to colors.
• That is, given an (x,y) pair, use a function to determine a color for the pixel at (x,y).
• Syntax: (image-compute *function* *width* *height*)
  • No exclamation point because we are not changing anything that alread exists; we are simply computing a new image.

Examples

• Color blends are the most natural thing to make.
• We can also make shapes by using conditionals.
• Given that colors are just numbers, we can also just do numeric computations and see what they give us.