Identification of Catalogued and Uncatalogued Classes.
Abstract
A technique to discriminate listed objects from unlisted ones is described. It is based on the principle of minimizing the probability of error in an identification process. The devised classifier is implemented as a threshold test. The classifier was applied to an aircraft identification problem. It was shown that the error of misclassifying catalogued targets as uncatalogued and vice versa can be made very small, while keeping a high probability of correct identification when the presence of a listed object is detected. The implementation of the developed scheme was shown to be simple and efficient. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1978
- Accession Number
- ADA064843
Entities
People
- Heng-cheng Lin
Organizations
- Ohio State University