Comics are a powerful medium for communicating information, which are especially useful in multi-language situations and for people with difficulty reading.

Part of the magic of comics is that the level of visual realism may range from the cartoon to the photographic. Simplified cartoon outlines and flat colors are often able to present even the most frightening concepts in a non-threatening and often humorous manner; classic example: airline safety cards.

I've been experimenting with different ways to simplify images for quite a while now. Back before full color displays and printers I often needed to make images look good with a limited number of colors.

Here is an example of an image processing technique that makes images look good on binary output. It combines dynamic thresholding (for the high-frequency edges) with quantization (for the low-frequency shadows and highlights). The effect is not as noisy-looking as error-diffusion, the common alternative, and decidedly comic.

As part of a research group at AT&T Labs, focused on how to help teenagers communicate, I came up with a way to simplify making comics by starting from videos. I implemented it as a Java video player, a video sampler, and a graphics editor with balloons that would reshape as you changed the text inside them and filters for simplifying the video images. You could mask out parts of images by painting on them. Kids could have fun examining their favorite movies, finding the individual frames that best represent an action, and adding their own dialog. The more adventurous could shoot their own videos to tell their own stories.

It was a lot of fun trying to convince managers at AT&T about the value of comics for communicating information. I was able to get this software released, but people could only download it by agreeing to a pretty restrictive license. Now, sadly, the software is no longer available. But here are some comics I made while I was working on it.