An investigation into the use of the FST classifier for ATR

Author

Connelly, Andrew ; Murray, Alan ; Renshaw, David

Author_Institution

Dept. of Electr. Eng., Edinburgh Univ., UK

fYear

1998

fDate

14-17 Sep 1998

Firstpage

381

Lastpage

386

Abstract

Automatic target recognition (ATR) is an area in which there has been significant research. One of the problems of ATR is classification confidence as objects may be mobile and change their poise. The paper describes the use of the feature space trajectory (FST) classifier to determine both the class and poise of objects. Investigations have been carried out using both synthetic and real data (long range infra-red) and results show that the FST classifier can provide additional valuable information to the ATR process. Future enhancements of the FST classifier are also discussed

Keywords

feature extraction; image classification; image segmentation; object recognition; automatic target recognition; classification confidence; feature space trajectory classifier; long range infrared data; Artificial neural networks; Data mining; Feature extraction; Image databases; Image edge detection; Image segmentation; Infrared imaging; Object recognition; Pattern recognition; Target recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control (ISIC), 1998. Held jointly with IEEE International Symposium on Computational Intelligence in Robotics and Automation (CIRA), Intelligent Systems and Semiotics (ISAS), Proceedings

Conference_Location

Gaithersburg, MD

ISSN

2158-9860

Print_ISBN

0-7803-4423-5

Type

conf

DOI

10.1109/ISIC.1998.713692

Filename

713692