This paper examines some of the reasons that tactical target recognition performance has not yet achieved a level necessary to satisfy the requirements for a totally autonomous target recognition system. In particular, some factors that adversely impact recognizer performance are identified and qualitatively analyzed to show how they impact performance. These factors include: training with inappropriate data sets and unknown target aspect angle. The problem of training and then estimating the performance of autonomous target recognition systems is also addressed. Commonly used techniques for training and performance estimation are appraised, including sample partioning and leaving-one-out estimation. Two issues are given particular attention: the multiclass nature of the target recognition problem and the requirement to recognize targets over a broad range of operating conditions.
Issues in automatic target recognition performance and performance estimation
Untersuchungen an Systemen zur automatischen Zielerfassung und Abschaetzung der Arbeitsweise
1982
5 Seiten, 8 Bilder, 2 Tabellen, 14 Quellen
Aufsatz (Konferenz)
Englisch
British Library Conference Proceedings | 1999
|Automatic target recognition performance losses in the presence of atmospheric and camera effects
British Library Online Contents | 2010
|