Self-Supervised Noise2Noise-Enhanced Denoising for Continuous-Scan Air-Plasma THz Spectroscopy
2026-08-17 • Machine Learning
Machine Learning
AI summaryⓘ
The authors worked on improving terahertz time-domain spectroscopy (THz-TDS), a technique used to capture very high-frequency signals, which normally requires many repeated measurements to get clear results because of noise and fluctuations. They developed a machine learning method that cleans up noisy data from just a single measurement, using a special neural network trained in two ways: one with clean examples and one without. Their combined approach makes it possible to get the same quality of data as averaging about five noisy measurements, speeding up the process without changing the hardware. This finding could help make THz-TDS faster and more efficient.
Terahertz time-domain spectroscopyAir-plasma generationBalanced air-biased coherent detectionSignal-to-noise ratioResidual U-NetNoise2Noise learningWiener filteringSelf-supervised learningContinuous-scanDenoising
Authors
Adam Umra, Oways Alsoloh, Oliver Nagy, Aydin Sezgin, Clara Saraceno
Abstract
Terahertz time-domain spectroscopy (THz-TDS) based on air-plasma generation and balanced air-biased coherent detection offers gap-free broadband coverage, but individual continuous-scan traces are strongly affected by pulse-to-pulse fluctuations and electronic noise. Reaching a useful signal-to-noise ratio therefore requires averaging multiple traces, which directly increases measurement time. We propose a learned denoising approach that recovers high-quality THz waveforms from as few as one complete continuous delay sweep, referred to here as a single-scan trace. A compact one-dimensional residual U-Net is trained using two complementary strategies: a reference-supervised baseline that maps individual noisy traces to long-average reference waveforms, and a Noise2Noise approach that learns from pairs of independently acquired noisy traces without requiring a clean training target. Averaging the predictions of both models reduces systematic bias and yields a trace-reduction factor of approximately $5.4\times$ at $K=1$, meaning that one denoised trace achieves the reconstruction accuracy of averaging approximately five raw traces. The Noise2Noise model alone achieves $4.9\times$, outperforming both the reference-supervised baseline ($4.6\times$) and classical Wiener filtering ($3.2\times$). These results show that self-supervised learning from repeated noisy measurements can support faster continuous-scan THz-TDS without hardware modification.