Fig. 4: Impact of Divisive Normalisation on Radio Frequency Interference (RFI) Detection in LOFAR Spectrograms.
From: Spiking neural networks for radio frequency interference detection in radio astronomy
a original spectrogram; b spectrogram after divisive normalisation; c latency-based inference on the original spectrogram; d latency-based inference on the normalised spectrogram showing a significantly more conservative flagging approach; e residual between original inference and mask; f residual between normalised inference and mask, and g expert-labelled RFI mask. While divisive normalisation improves clarity, the SNN output still exhibits significant noise compared to expert labelling. This is most apparent in the residual plots where in both cases key features are missing from inference.
