Computer vision algorithms are not general-purpose and rely on specific assumptions about the video input, while human vision is adaptable and can interpret a wide range of visual stimuli. Unlike humans, computers cannot inherently understand or extract semantic information from visual data without additional programming. Additionally, computer vision systems may struggle with ambiguous or poorly defined scenes, whereas humans can often infer meaning from context.
Techniques to enhance computer vision include controlled illumination to improve contrast, using brightness thresholding to distinguish objects based on their brightness, and employing background subtraction to isolate moving objects. Additionally, surface treatments like high-contrast paints can make objects more detectable. Simple object tracking algorithms can also be implemented to follow specific features, such as the brightest pixel in a video frame.
Computer vision’s tracking and surveillance capabilities enable interactive art to engage audiences by responding to their movements and expressions, creating immersive experiences. However, this capacity raises ethical concerns regarding privacy and consent, as participants may be unaware of being monitored. Artists must balance the innovative use of tracking technology with responsible practices to ensure a respectful and enjoyable interaction.