CFARnet: deep learning for target detection with constant false alarm rate
This work addresses the CFAR constraint for target detection in practical applications where classical methods are computationally expensive or data-driven, offering a novel deep learning approach that is incremental by building on classical hypothesis testing.
The paper tackles the problem of target detection with a constant false alarm rate (CFAR) constraint, introducing CFARnet, a deep learning framework that approximates a CFAR-constrained Bayes optimal detector, which is shown to be asymptotically equivalent to the classical generalized likelihood ratio test (GLRT). Experiments demonstrate that CFARnet enables a flexible tradeoff between CFAR and accuracy in various target detection settings.
We consider the problem of target detection with a constant false alarm rate (CFAR). This constraint is crucial in many practical applications and is a standard requirement in classical composite hypothesis testing. In settings where classical approaches are computationally expensive or where only data samples are given, machine learning methodologies are advantageous. CFAR is less understood in these settings. To close this gap, we introduce a framework of CFAR constrained detectors. Theoretically, we prove that a CFAR constrained Bayes optimal detector is asymptotically equivalent to the classical generalized likelihood ratio test (GLRT). Practically, we develop a deep learning framework for fitting neural networks that approximate it. Experiments of target detection in different setting demonstrate that the proposed CFARnet allows a flexible tradeoff between CFAR and accuracy.