Pattern recognition

Pattern recognition software mimics visual perception by searching an image for a specific pattern. This field of computer science was first developed in the 1950s at academic institutions such as the M.I.T. A.I. Lab. It is commonly used in Computer Aided Machinery[?] (CAM), robotics, computer vision, and machine vision.

Pattern recognition includes both Optical Character Recognition (OCR) and the identification of outlines, facial features, and textures.

Pattern recognition requires a video camera or some other scanning device to capture images. These images are then used as inputs to computer processing software that can read text, identify a face, or specify the location of any object the system has been trained to recognize.

The process of pattern recognition may be described as follows:

1. A digital video camera images an area in which the object to be recognized is expected to appear
2. A "framegrabber" computer card captures an image from the camera's video output and stores it in computer memory as an individual digital image
3. An algorithm detects edges and/or other features of interest in the captured image
4. Another algorithm analyzes the processed image and determines if a specific pattern exists or not

Hidden edges can be extrapolated from 2D data to generate 3D datasets for robotics control in a 6 Degrees of Freedom[?] (6 DOF) environment. Dual cameras for binocular vision are also used.

See also:

• artificial intelligence
• computer vision
• machine vision
• optical character recognition
• barcode
• music information retrieval[?]

Reference

• Richard O. Duda, Peter E. Hart, David G. Stork (2000) Pattern classification (2nd edition) , Wiley, ISBN 0471056693