Title :
Sensitivity analysis of ATR algorithms to scene distortions: the key to performance evaluation of automatic target recognition systems
Author :
Mohd, Maqsood A.
Author_Institution :
TEAS Group, Sverdrup Technol. Inc., Eglin AFB, FL, USA
Abstract :
This paper presents a methodology to characterize the automatic target recognition (ATR) algorithm performance when it is subjected to geometrical scene distortions. The methodology presented is a parametric variation technique whereby perturbations in sensor parameters are made and the target probability of detection and location determined. The generic methodology developed here can serve as a very useful tool in the test and evaluation of ATR systems used in industrial, scientific, commercial, medical, civil defence, and national defence applications
Keywords :
image recognition; image sequences; probability; ATR algorithms; automatic target recognition algorithm; automatic target recognition systems; detection probability; location probability; parametric variation technique; performance evaluation; sensitivity analysis; sensor parameters; Gold; Image recognition; Image sensors; Infrared image sensors; Layout; Object detection; Performance evaluation; Sensitivity analysis; Target recognition; Testing;
Conference_Titel :
Southeastcon '94. Creative Technology Transfer - A Global Affair., Proceedings of the 1994 IEEE
Conference_Location :
Miami, FL
Print_ISBN :
0-7803-1797-1
DOI :
10.1109/SECON.1994.324266
