Self-supervised denoising improves vocal tract MRI video quality
Blind Video Denoising for Vocal Tract rtMRI using a Self-Supervised Ensemble
Computer Vision and Pattern Recognition
Summary
Real-time MRI videos of the vocal tract can show how speech sounds are made, but the videos are noisy and hard to see clearly. Since there are no noise-free examples to teach from, the authors developed a new method that learns to reduce noise without clean references. They combine different techniques to keep important details while removing noise. Their tests show this method works better than other noise-reduction techniques, making the videos clearer and more detailed.
What this means in practice
- •For speech therapists: Improve the clarity of vocal tract MRI videos used to assess speech articulation without needing clean reference images.
- •For radiology technologists: Enhance noisy real-time MRI scans to provide clearer images for faster clinical interpretation during speech studies.
Authors
Julian Andres Hernandez-Potes, Tomás Arias-Vergara, Daiqi Liu, Andreas Maier, Paula Andrea Perez-Toro
Abstract
Real-time magnetic resonance imaging of the vocal tract makes speech articulation directly observable, but the trade-off between temporal resolution and signal quality introduces substantial noise and obscures fine articulatory detail. Since clean references cannot be acquired for a moving vocal tract, supervised denoising approaches are not applicable in this setting. We propose a self-supervised denoising framework that combines spatiotemporal super-resolution learning with pseudo-supervision derived from a variance-stabilized non-local means filter. A dual-model fusion strategy balances noise suppression and preservation of high-frequency articulatory structure. Experiments on the USC-75 multi-speaker dataset under natural noise conditions show that the ensemble raises the signal-to-noise ratio by 11.9 dB on unseen subjects while attaining more high-frequency articulatory content in the frames. Among tuned implementations of anisotropic diffusion, block-matching filtering, non-local means and pre-smoothed non-local means, every configuration that achieves comparable noise suppression does so at the cost of high-frequency content or of structural fidelity. Under severe added Gaussian, Rayleigh and Rician noise, peak signal-to-noise ratio improves by 7.6 to 11.8 dB and structural similarity roughly triples. An ablation shows that the variance-stabilized pseudo-supervision is an essential element for denoising, and that the effect is attributable to the hard thresholding contained in the stabilization transform.