August 12, 2010

Producing Better Facial Composites for Police Use

World of Warcraft players call them 'Toons', Nintendo fans probably have a 'Mii' but the rest of us are more familiar with the term 'avatar'

These digital representations of one’s self are commonplace within the video game industry. Players customize the physical features of their avatar as they see fit. The game developers probably expect players to render a true-to-life self-portrait but what often happens is quite the opposite. Players choose to create an avatar they feel is a ‘better’ representation of their self. They exaggerate, enhance or completely re-envision their self-image. Customizing them is a significant part of the overall experience.

I read an online news article this morning that featured an image released by the Michigan State Police. The image was a facial composite of a man suspected in connection with a three-state serial stabbing spree. Let’s take a look at the sketch and a photograph of the arrest suspect.

Now, let’s compare the achievements of the sketch vs. the photograph. The sketch simply depicts a male with facial fair and a hat. The sketch doesn't depict an ethnicity, eye color, hair color or any of the correct facial features.

It's about time we ditched traditional police sketch artist. According to Wikipedia, “the FBI maintains that hand-drawing is still the correct method for constructing a facial composite.” What if, instead, they utilized the system of avatar creation from Video Games? Victims/Witnesses would digitally construct features of criminals without having to wait for the render by the sketch artist. These digital composites should not be as simple as Nintendo Miis, yet still more complex than XBOX 360 avatars. The user could choose which features to start with and revisit/revise their creation throughout the process.

My name is Verne Troyer, I live in Hollywood, California, and I’m a mage! A master of the arcade.