Computer vision began as a distinct discipline in 1966 when an undergraduate summer project attached a camera to a computer. This early experiment marked the start of artificial intelligence research at universities focused on describing what machines saw.
Current applications include facial matching for mobile phone facelock, optical character recognition for text encoding, and emotion recognition systems. Medical image analysis detects diseases in cardiology and radiology while agricultural processes use machine vision to sort foodstuff automatically.
Most computer vision systems rely on image sensors designed using quantum physics to detect electromagnetic radiation. Light interacts with surfaces according to principles explained by solid-state physics which governs optics in imaging systems.
Industry held a market size of US$5.22 billion in 2024 while medicine reached US$2.6 billion and military operations totaled US$996.2 million. Wafer inspection measures every single wafer for inaccuracies or defects to prevent unusable computer chips from reaching markets.
Recent work has seen deep learning algorithms surpass prior methods on benchmark datasets for tasks ranging from classification to optical flow. Current best algorithms based on convolutional neural networks now approach human-level accuracy on these tests.