Multinomial feature matching for out-of-distribution detection in synthetic aperture radar
Christopher W. Pitts,
Devin White,
Trilce Estrada,
and Gruia-Catalin Roman
In Automatic Target Recognition XXXVI
2026
National Harbor, Maryland, USA
Received Best Paper Award
Out-of-distribution (OOD) detection is an important part
of automatic target recognition (ATR) systems. The
capability to reject unknown classes improves
reliability and trust in an ATR, and permits the use
of otherwise closed-set classifiers where open-set
recognition is necessary. In this paper we present
multinomial feature matching (MFM), a method for
detecting OOD data in the latent feature space of
neural classifiers, and apply it to an
EfficientNet-B7 model trained on the SAMPLE+
dataset. We show that MFM has efficient and
low-overhead runtime characteristics, and that it
exhibits a high level of performance when applied to
OOD targets in SAMPLE+ and other SAR datasets,
including both vehicle targets and clutter. MFM
achieves a state-of-the-art area under the receiver
operating characteristic (ROC) curve (AUROC) score
and false positive rate (FPR) at a 95% true
positive rate (TPR) (FPR@95) benchmark on the
dataset, outperforming both mainstay benchmarks in
OOD detection and contemporary work in the field.