Computer Vision: A Control Freak’s Nightmare (But Also Their Dream)

(Response to “Computer Vision for Artists and Designers: Pedagogic Tools and Techniques for Novice Programmers” by Golan Levin)

Levin’s article about computer vision succeeds in describing the field through its many dimensions: he discusses its history, main obstacles, common (and useful) techniques, and popular uses (among other aspects). The text is thus informative, instructive, and a reflection of what computer vision is capable of now and its potential for the future.

Two of Levin’s points struck me the most. None of them were “news” (I was familiar with these facts beforehand), but it was their very obviousness (which, by the way, is a word… I had to check) that made me reflect on them. The first is that we are fascinated by the idea of surveillance. The second is that computer vision is, as Levin puts it, “computationally ‘opaque’… [it] contains no intrinsic semantic or symbolic information.”

I’ll begin with the latter. In developing my own computer vision project for this class, I couldn’t stop thinking about how complicated it is to control how the program works and the users’ experience with it. As Levin correctly explains, computer vision can only analyze pixel by pixel and then interpret the information of these individual units. In this sense, control is particularly challenging. However, because so little can be controlled by the basic algorithms we’re using (in comparison to coding only with text, as Levin states), the “outside” (what’s on screen) often has to be a very controlled environment – light levels, colors, motion, etc. There are numerous conditions that can greatly affect the functioning of the program, which the computer can´t deal with on its own.

In terms of surveillance, I wonder why the camera inspires so many works about this topic. It’s certainly related to our awareness of the fact that our degree of privacy is often an illusion (in this age, technology has evolved to track almost anyone). The camera has the ability to serve as a second pair of eyes, and “see” what we couldn’t see without it. We can act as those who carry out the surveillance, which puts us in a position of power, but we also realize that others can do the same to us, which is worrisome.