Learning to Detect with Constant False Alarm Rate
This addresses a critical limitation in machine learning-based hypothesis testing for applications needing reliable false alarm control, though it is incremental as it builds on existing learned detectors.
The paper tackles the problem of learned detectors lacking a constant false alarm rate (CFAR) required in applications like target detection, and proposes adding a loss term to promote similar null hypothesis distributions, resulting in near CFAR detectors with similar accuracy to competitors.
We consider the use of machine learning for hypothesis testing with an emphasis on target detection. Classical model-based solutions rely on comparing likelihoods. These are sensitive to imperfect models and are often computationally expensive. In contrast, data-driven machine learning is often more robust and yields classifiers with fixed computational complexity. Learned detectors usually provide high accuracy with low complexity but do not have a constant false alarm rate (CFAR) as required in many applications. To close this gap, we propose to add a term to the loss function that promotes similar distributions of the detector under any null hypothesis scenario. Experiments show that our approach leads to near CFAR detectors with similar accuracy as their competitors.