It is really hard to outdo interactions that map the objects existing in our physical reality onto our screens for us to warp and play around with, so it was not surprising to me that computer vision techniques would be popular in the production of interactive art. One of the things that surprised me, however, was how early the adoption of computer vision principles – as early as the late 1960s– occurred. In addition to introducing a brief history of computer vision in interactive art, the reading also delineates simple computer vision algorithms that can be exploited by digital creators. Moreover, while the article acknowledges the complexity of the field of computer vision, it introduces the audience to a wide range of multimedia tools that make computer vision more accessible to beginners who are not familiar with machine vision.
Given the power that computer vision grants creators in being able to recreate, manipulate, and reshape physical reality, artworks created with computer vision techniques have the potential to represent the multiple layers of complexity that exist in the world around us. It was interesting – and shocking at times – to see how different sociopolitical themes can be expressed using the power of computer vision. This is demonstrated best by the pieces created on surveillance that the article mentions. The idea of using belts controlled by a tracking system that rotates according to the motion of the audience to highlight surveillance by an “empty crowd” in Rafael Lozano-Hemmer’s installation, Standards and Double Standards, exemplifies how well computer vision can “metaphorize” our physical reality. However, additional considerations, and potentially greater risks, come with more capability. An example of this is the Suicide Box, which captured video footage of suicide jumpers from the Golden Gate Bridge. It was jarring for me to see a tragic phenomenon turned into a piece of “art” and was a testimony to how computer vision technology, like any tool with great computational power, must be accompanied by a recognition of the ethical implications of its use.
I also enjoyed how the reading emphasized the importance of using computer vision techniques that are compatible with the surrounding physical environment as no particular algorithm will be truly generalizable to all video/image formats. This may take the form of making the physical environment conducive to optimizing the workings particular algorithm or choosing an algorithm that is best suited for a given environment. I loved learning about the different techniques that designers employ to achieve this as it effectively extends the idea of Human-Centered Interaction to Environment-Centered Interaction and truly captures the unleashing of interactivity potential that computer vision affords digital creators.