![[2602.22431] mmWave Radar Aware Dual-Conditioned GAN for Speech Reconstruction of Signals With Low SNR](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22431] mmWave Radar Aware Dual-Conditioned GAN for Speech Reconstruction of Signals With Low SNR
Summary
This article presents a novel approach using a Dual-Conditioned Generative Adversarial Network (GAN) for reconstructing speech signals captured by mmWave radar under low signal-to-noise ratios.
Why It Matters
The research addresses the challenge of reconstructing intelligible speech from noisy mmWave radar signals, which is crucial for applications in surveillance, communication, and human-computer interaction. Improving speech clarity in such conditions can enhance user experience and expand the utility of radar technologies.
Key Takeaways
- Introduces a two-stage speech reconstruction pipeline using RAD-GAN.
- Demonstrates improved performance over existing methods in low SNR conditions.
- Utilizes a Multi-Mel Discriminator and Residual Fusion Gate for enhanced processing.
- Trained on a limited dataset without pre-trained modules or data augmentations.
- Provides online audio examples to showcase the effectiveness of the proposed method.
Computer Science > Sound arXiv:2602.22431 (cs) [Submitted on 25 Feb 2026] Title:mmWave Radar Aware Dual-Conditioned GAN for Speech Reconstruction of Signals With Low SNR Authors:Jash Karani, Adithya Chittem, Deepan Roy, Sandeep Joshi View a PDF of the paper titled mmWave Radar Aware Dual-Conditioned GAN for Speech Reconstruction of Signals With Low SNR, by Jash Karani and 3 other authors View PDF HTML (experimental) Abstract:Millimeter-wave (mmWave) radar captures are band-limited and noisy, making for difficult reconstruction of intelligible full-bandwidth speech. In this work, we propose a two-stage speech reconstruction pipeline for mmWave using a Radar-Aware Dual-conditioned Generative Adversarial Network (RAD-GAN), which is capable of performing bandwidth extension on signals with low signal-to-noise ratios (-5 dB to -1 dB), captured through glass walls. We propose an mmWave-tailored Multi-Mel Discriminator (MMD) and a Residual Fusion Gate (RFG) to enhance the generator input to process multiple conditioning channels. The proposed two-stage pipeline involves pretraining the model on synthetically clipped clean speech and finetuning on fused mel spectrograms generated by the RFG. We empirically show that the proposed method, trained on a limited dataset, with no pre-trained modules, and no data augmentations, outperformed state-of-the-art approaches for this specific task. Audio examples of RAD-GAN are available online at this https URL. Comments: Subjects: Sound (cs.S...