computer vision

Computer vision has been one of the major areas of AI since its earliest days. Typical image processing pipelines include appliying low-level filters, followed by edge detection followed by some sort of object recognition and eventually more semantic understanding. Various forms of deep neural network are now used extensively, where these distinctions between stages are effectively buried within the network. However, understanding stages of vision processing continues to be important, for example, providing the insight behind convolutional neural networks.