Image Arithmetic
Final Exam

Well, has all this IA detail soaked in a little? Here's your chance to see if you can predict what's going to happen when I apply image arithmetic to a couple of images. Don't feel bad if you guess wrong — IA is primarily trial and error, and you usually just don't know what you're going to get until you try it. But I hope you've gotten some feel for which IA functions are best to try for whatever effect you're trying for.

Unless indicated differently, all aritimetic is clipped with divisor=1 and bias=0.

Here are the two images we'll be using:

Average

Let's start with the "average" method with a divisor of 1. This is equivalent to the "add" method with a divisor of 2. Before you peek, take a guess what the result will look like and whether the result will be a pleasing one. Look at both of the originals before you peek at the result.

Original #1
Original #2
Average

The value of each pixel's three RGB components is exactly halfway between the corresponding components of the two images. Is this about what you expected?

Lightest and Darkest

Let's try the "lightest" and "darkest" methods. Again, try to visualize the results before you peek. Do you think "lightest" or "darkest" will produce the best results for this image?

Original #1
Original #2
Lightest
Darkest

Multiply

Let's try "multiply". Remember that you'll usually want a large divisor. I've used a divisor of 150 for this example. Any ideas what this'll look like?

Original #1
Original #2
Multiply

OR and AND

Time for the OR and AND methods. Remember the general features of these - each pixel in an OR'd result will be the same or lighter than in each original, and each pixel in an AND'd result will be the same or darker than in each original. How do you think the results will appear? Will OR or AND produce a better result for these images?

Original #1
Original #2
OR
AND

Subtract and Difference

Now let's tackle the "subtract" and "difference" methods. There are three images here. The first is image 1 (the lighthouse) subtracted from the other. The second is image 2 subtracted from the lighthouse image. The third is the difference method, which uses the absolute value of each subtraction result and is not dependent on which image is subtracted from which.

Any wild guesses how these will look?

Original #1
Original #2
#1 - #2
#2 - #1
Difference

Don't feel bad if you found that you couldn't predict what these results were going to look like. That's just the way it is with image arithmetic — you may have some idea of the effect you want to achieve, but it usually takes quite a lot of trial-and-error and a lot of patience to achieve that effect.