Accession Number:

ADA011675

Title:

Theory and Applications of Image Pattern Recognition.

Descriptive Note:

Interim technical rept.,

Corporate Author:

SOUTHEASTERN MASSACHUSETTS UNIV NORTH DARTMOUTH DEPT OF ELECTRICAL ENGINEERING

Personal Author(s):

Report Date:

1975-02-24

Pagination or Media Count:

81.0

Abstract:

In recent years, there has been increasingly great demand for automatic recognition of imagery patterns which arise in biomedical, space, military, security and vast number of other applications. As the computer hardware cost decreases rapidly, automatic image recognition will soon become a reality in many of these applications. Automatic image recognition is also a major step toward designing intelligent machines. In this report, a unified theory of image pattern recognition is developed which includes pattern representation, preprocessing, feature extraction and pattern classification. Computer results from aerial photograph studies are used for illustrative purposes. The theory should be useful for a wide range of imagery patterns. The problem of computational complexity which should be the primary consideration in recognition system design is carefully examined. As the imagery patterns are complicated in nature, it is recommended that both statistical and syntactic methods in pattern recognition should be used to arrive at the best solution for the specific recognition task at hand. Extensive bibliography at the end of the report provides a fairly complete view of the past efforts, approaches and applications of image pattern recognition.

Subject Categories:

  • Computer Programming and Software

Distribution Statement:

APPROVED FOR PUBLIC RELEASE