ON THE RECOGNITION OF SIGNAL PATTERNS IN NOISE
Abstract
IN THE APPLICATION OF STATISTICAL DECISION THEORY TO THE RECOGNITION OF TWO-DIMENSIONAL PATTERNS IN NOISE, THE DECISION AS TO WHICH OF A SET OF PATTERNS IS PRESENT INVOLVES THE CALCULATIONS OF A LIKELIHOOD FUNCTION FOR EACH OF THE POSSIBLE SIGNALS. In this paper, systems are developed for the calculation of these functions when the signal patterns are perturbed by multiplicative and additive noise. The resulting realizations ARE SHOWN TO REDUCE TO TWO-DIMENSIONAL MATCHED FILTERS. In the development, a useful orthogonal expansion of two-variable random processes is derived. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1961
- Accession Number
- AD0260304
Entities
People
- Jack K. Wolf
- John B. Thomas
Organizations
- Rome Laboratory