<<Up     Contents

Pattern recognition

The visual cortex is excellent in 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:

Reference

Glasglow.com