Tanumoy Saha
Tanumoy Saha is a Research Scientist at the Deutsches Zentrum für Astrophysik (DZA), Germany.
He holds a PhD in Physics and has a background in physical modelling, numerical simulations, and machine learning. His current research focuses on developing AI-driven methods for pulsar detection and analysis, combining deep learning, signal processing, and computer vision to improve pulsar search pipelines.
Session
We present an ongoing investigation of a non-learnable approach for pulsar-signal detection using handcrafted 2D line filters. The method applies Gaussian ridge filters across multiple scales and orientations to identify the characteristic dispersed structures of pulsars in dynamic spectra. Small random perturbations are introduced into the filters, creating a noise-augmented filter bank whose responses are subsequently aggregated to reinforce consistent signal structures while suppressing uncorrelated fluctuations.
The filter responses are processed through a sequence of sigmoid activation, nonlinear transformation, multi-scale summation, binarization, orientation alignment, and peak-prominence selection. This adaptive processing identifies orientations that provide consistent evidence for a pulsar signal without relying on learned convolutional kernels.
The current results on the B0531+21 dataset demonstrate the potential of this handcrafted filtering approach, with an overall F1-score above 90%.