Spring 2009
22C:251 Visualization
Assignment 1
Due: Tuesday, February 24th at 11:59pm
Goal: Compile a program using the Visualization Toolkit. Explore various coloring schemes for visualizing univariate
data in 2D images.
Problem 1 (10 points): Compile the Cone example shown in class. The code is posted online. Rotate, translate, or
scale the cone some random amount and post an image on a publically-accessible webpage you create to display your
results.
•You can use any utility you want to capture screenshots. (Good ways include using GIMP (File→Acquire→Screen Shot) or,
“PrintScreen” + “Paste” inside MS Paint in Windows.)
•You must send me an e-mail with the address of your webpage.
•I plan to link all web pages from the class homepage, unless you explicitly ask me not to.
Problem 2 (40 points): Create a program that loads an image from a file and visualizes it using the following
techniques. You must test these techniques on the following images (see webpage): MR knee.png,CT chest.png,
usincome.png,flow density.png, and rose 512.png. I encourage you to write your code so that it loads an image
specified on the command line. This way you will not need to modify your code independently for each image.
•Part A: Convert the images to a luminance scale (Hint: vtkImageLuminance), then create a lookup table (Hint:
vtkImageMapToColors + vtkLookupTable) that classifies the image into 3, 5, 7, or 9 grayscale categories. Use
the following three methods for this classification:
1. Equal intervals. (i.e., for 3 categories: gray values [0..85] are colored 42, [86..170] are colored 127,
[171..255] are colored 212)
2. Separate quantiles. (i.e., for 3 categories: the dimmest 1
3of pixels are colored one value, the middle 1
3
another, and the top 1
3another)
3. Mean + Standard Deviation. (i.e., for 3 categories: within 1 standard deviation of the mean are colored
the middle gray, less than 1 standard deviation below the mean are colored a darker gray, greater than 1
standard deviation above the mean a lighter gray)
Note: For the quantiles and mean/standard deviations, the appropriate values will be posted online so you need not compute
them. On your webpage, you must post 5 images for each classification method (one for each of the input PNG files). This
image can either have 3, 5, 7, or 9 categores but must be the one you think conveys the most information for that image and
classification scheme.
•Part B: Implement the following color scales (Hint: vtkImageMapToColors +vtkLookupTable). For each input
image and color scale, test it with the original 256-color grayscale image and the three grayscale categories
from Part A. Of these four tests, put the one you think conveys the most information on your webpage.
1. A perceptually linear gray scale. (See webpage)
2. A single-hue color scale (e.g., just magenta, red, or blue of varying saturations)
3. A two-hue opponent color scale (e.g., yellow-blue, red-green, red-cyan)
4. A rainbow scale (e.g., (1,0,0)→(1,1,0)→(0,1,0)→(0,1,1)→(0,0,1)→(1,0,1))
REMEMBER: Send me an e-mail listing your web address where I can go to see your images.
