A fundamental problem facing the designers of automatic target recognition (ATR) systems is how to deal with out-of-library or non-registered targets. This research extends a mathematical programming framework that selects the optimal classifier ensemble and fusion method across multiple decision thresholds subject to classifier performance constraints. The extended formulation includes treatment of exemplars from target classes on which the ATR system is not trained (non-registered targets).
Further, a multivariate Gaussian hidden Markov model (HMM) is developed and applied using real world synthetic aperture radar (SAR) data comprised of ten registered and five non-registered target classes. The framework is exercised in an experimental design across classifier fusion methods, prior probabilities of targets and non-targets, correlation between multiple sensor looks, and levels of target pose estimation error.
[1] USAF (2000), Air Warfare, AFDD 2–1. Dept of the Air Force, Washington DC.
2.
[2] JFCOM (2001), Capstone Requirements Document: Combat Identification. United States Joint Forces Command (J-8), Norfolk VA.
3.
[3] Sadowski, C. (2005), Combat identification: Where are we?, AFIT OPER 596 Guest Lecture.
4.
[4] USAF (1999), Intelligence, Surveillance, and Reconnaissance Operations, AFDD 2–5.2. Dept of the Air Force, Washington DC.
5.
[5] Roli, F. (2003), Fusion of multiple pattern classifiers.In 8th National Conference of the Italian Association of Artificial Intelligence.
6.
[6] Fumera, G., Roli, F., and Giacinto, G. (2000), Reject option with multiple thresholds.Pattern Recognition, 33: 2099–2101.
7.
[7] Fumera, G., Roli, F., and Vernazza, G. (2001), A method for error rejection in multiple classifier systems. In Proceedings of the 11th International Conf on Image Analysis and Processing, pp. 454–458.
8.
[8] Fumera, G., Pillai, I., and Roli, F. (2004), A two-state classifier with reject option for text categorisation.In 5th Int. Workshop on Statistical Techniques in Pattern Recognition (SPR 2004), vol. 3138, pp. 771–779.
9.
[9] Fumera, G., and Roli, F. (2004), Analysis of error-reject trade-off in linearly combined multiple classifiers.Pattern Recognition, 37(6): 1245–1265.
10.
[10] Laine, T. (2005), Optimization of Automatic Target Recognition with a Reject Option Using Fusion and Correlated Sensor Data. Ph.D. dissertation, Air Force Institute of Technology, Wright-Patterson AFB, OH.
11.
[11] Laine, T. and Bauer, K. (2006), A mathematical framework to optimize ATR systems. Computers and Operations Research, Special Issue on OR Applications in the Military and in Counter-Terrorism, accepted Apr 2006, to appear.
12.
[12] Alsing, S., Bauer, K., and Miller, J. (2002), A multinomial selection procedure for evaluating pattern recognition algorithms.Pattern Recognition, 35: 2397–2412.
13.
[13] Alsing, S., Bauer, K., and Oxley, M. (2002), Convergence of receiver operating characteristic curves and the performance of artificial neural networks.International Journal of Smart Engineering Systems Design, 4: 133–145.
14.
[14] Alsing, S. (2000), The Evaluation of Competing Classifiers. Ph.D. dissertation, Air Force Institute of Technology, Wright-Patterson AFB, OH.
15.
[15] Chow, C. (1970), On optimum rejection error and reject tradeoff.IEEE Transactions on Information Theory, IT-16: 41–46.
16.
[16] Audet, C., and Dennis, J. (2000), Pattern search algorithms for mixed variable programming.SIAM Journal of Optimization, 11(3): 573–594.
17.
[17] Albrecht, T. (2005), Combat Identification with Sequential Observations, Rejection Option, and Out-of-Library Targets. Ph.D. dissertation, Air Force Institute of Technology, Wright-Patterson AFB, OH.
18.
[18] AFRL (2005), Sensor Data Management System. Note: research used DCS data set 2005, available from AFRL but may not be posted on website. Air Force Research Laboratory, Wright-Patterson AFB, OH. https://www.sdms.afrl.af.mil/main.php.
19.
[19] Williams, R., Gross, D., Palomino, A., Westerkamp, J., and Wardell, D. (1998), 1d HRR data analysis and ATR assessment. In Proceedings of SPIE, vol. 3370, pp. 588–599.
20.
[20] Fielding, K. (1994), Spatio-Temporal Pattern Recognition using Hidden Markov Models. Ph.D. dissertation, Air Force Institute of Technology, Wright-Patterson AFB, OH.
21.
[21] Fielding, K., and Ruck, D. (1995), Spatio-termporal pattern recognition using hidden markov models.IEEE Transactions on Aerospace and Electronic Systems, 31(4): 1292–1300.
22.
[22] Rabiner, L. (1989), A tutorial on hidden markov models and selected applications in speech recognition. Proceedings of the IEEE, 77: 257–285.
23.
[23] Meyer, G. (2003), Classification of Radar Targets Using Invariant Features. Ph.D. dissertation, Air Force Institute of Technology, Wright-Patterson AFB, OH.
24.
[24] Duda, R., Hart, P., and Stork, D. (2001), Pattern Classification.John Wiley and Sons, New York, NY, second ed.
