Pattern recognition - Definition and Overview

For the William Gibson novel, see: Pattern Recognition (novel).

Pattern recognition (also known as classification or pattern classification) is a field within the area of computer science and can be defined as "the act of taking in raw data and taking an action based on the category of the data" [1]. It uses methods from statistics, machine learning and other areas.

Typical applications are automatic speech recognition, classification of text into several categories (e.g. spam/non-spam email messages), the automatic recognition of handwritten postal codes on postal envelopes, or the automatic recognition of images of human faces. The last three examples form the subtopic image analysis of pattern recognition that deals with digital images as input to pattern recognition systems.

Pattern recognition techniques

Pattern recognition is typically an intermediate step in a longer process. These steps generally are acquisition of the data (image, sound, text, etc.) to be classified, preprocessing to remove noise or normalize the data in some way (image processing, stemming text, etc.), computing features, classification and finally post processing based upon the recognized class and the confidence level.

Pattern recognition itself is primarily concerned with the classification step. In some cases, such as in neural networks or genetic algorithms, feature selection and extraction may also be partially or fully automated.

While there are many methods for classification, they are all mathematically similar in that they all partition a multi-dimensional feature space into multi-dimensional volumes. The classification is generally based upon how close to the center of the training volume a particular sample falls. Of course, the model may be expressed in a way that you might not naturally associate with a multi-dimensional feature space. For example, a neural network initially doesn't seem to be doing this at all. There may also be differences in that some models give confidence levels, while others do not. Neural networks and genetic algorithms may have advantages over approaches requiring hand coding of features if choosing good features is difficult in a particular domain.

• Nearest Neighbor (Pattern Recognition)
• Neural Networks
• Hidden Markov Models
• Bayesian networks
• Genetic Algorithms
• Support vector machine

Application domains

• computer vision
• machine vision
• medical image analysis
• optical character recognition
• credit scoring

See also:

• artificial intelligence
• machine learning
• barcode
• information retrieval
• color histograms

References

1. Richard O. Duda, Peter E. Hart, David G. Stork (2001) Pattern classification (2nd edition), Wiley, New York, ISBN 0471056693.
2. J. Schuermann: Pattern Classification: A Unified View of Statistical and Neural Approaches, Wiley&Sons, 1996, ISBN 0471135348