作者: Laurence C. Lambert
DOI:
关键词:
摘要: A machine is disclosed that capable of locating human faces in video scenes with random content within two minutes, and recognizing the it locates. The uses images obtained from a camera insensitive to variations brightness, scale, focus, operates without any intervention or inputs. When motion detection feature included, (one its options), location recognition events occur less than 1 minute. One embodiment system uses: camera, Micro-Vax computer, an analog-go-digital A/D converter, hard copy print out process using original computer program locate identify them. In operation, converts into analog electrical signal, which converted digital forwarded computer. performs pattern algorithm search for facial components, gestalt face, compare gestalt-face's detected characteristics stored set known faces, face thereby.