STATISTICAL IMAGE RECOGNITION CONCEPT (OSIRIS).
Abstract
The initial process in a system for the machine recognition of imagery is an operation which produces a model or transformation which is characteristic of a particular image or class of images. This model, or transformation, is subsequently processed in a machine to produce a signal which indicates that the input image is, or is not, to some degree of probability, a particular example of a desired image or class of images. An initial PsubF(d) transformation of optical imagery was investigated at some length. Analytical expressions are obtained for the PsubF(d) transformations of rectangular figures and circles. A machine program was devised which produced PsubF(d) transformations for a more general universe of imagery. The PsubF(d) transformations thus produced are subjected to an analysis which shows the transformation to be a valid technique for the recognition of imagery both as to class membership and discrimination between images of the same class. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 29, 1962
- Accession Number
- AD0634514
Entities
People
- G. R. Tenery
Organizations
- Vought