Reading Levin’s “Computer Vision for Artists and Designers” made me realize how much I take my own sight for granted and got me questioning what it means for a computer to “see.” I think I always imagined it as something similar to human vision? just more technical with lots of codes and numbers and letters. This reading though, made it clear that computer vision is not actually vision in the way we experience it. While we humans automatically recognize faces, objects, details, and context without consciously thinking about it, a computer just receives pixel values without understanding the meaning since it only processes numbers and the differences between frames. That really stood out to me because it shows how much interpretation we naturally do as humans without noticing it.
Another important difference is that human vision is flexible while computer vision is mostly dependent on certain conditions, for example if the lighting or background changes a little, the computer vision system can fail completely. The reading emphasized how algorithms like frame differencing or background subtraction rely on stable environments and clear contrasts which made me realize that computer vision is not just about writing code but it’s also about carefully designing the physical environment so that the system can succeed. There are things like using a strong contrast or controlled lighting or also reflective materials can improve the tracking of the computer vision. This also allows the artist to be responsible for preparing the place or area so the computer can see clearly.
Overall, this reading made me see computer vision through a more structured system that has its own limitations. It also made me think about how these tracking systems can feel like surveillance, which gives the artist power to make the audience feel observed or even part of the piece in a more intense way. As someone working in interactive media, this makes me more aware that designing with computer vision is not just about making something work, but about understanding what the system can actually perceive and what it cannot.