# Speech enhancement deep-learning architecture for efficient edge processing

arXiv: https://arxiv.org/abs/2405.16834 · HTML: https://mlsanity.com/paper/2405.16834/

## Key facts

- Cited by 1 papers in the map (data through 19 Oct 2026).
- Placed in Speech enhancement and noise suppression, one of 11 directions in Speech and audio language models.
- Builds on 21 of its 32 references that are in the map.
- First posted on arXiv 27 May 2024.

## Placed in

- Speech and audio language models / [Speech enhancement and noise suppression](https://mlsanity.com/speech/speech-enhancement-and-noise-suppression/)

## Builds on (most cited first, 21 of 21)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification](https://mlsanity.com/paper/2005.07143/) (2020)
- [Res2Net: A New Multi-scale Backbone Architecture](https://mlsanity.com/paper/1904.01169/) (2019)
- [DCCRN: Deep Complex Convolution Recurrent Network for Phase-Aware Speech Enhancement](https://mlsanity.com/paper/2008.00264/) (2020)
- [Real Time Speech Enhancement in the Waveform Domain](https://mlsanity.com/paper/2006.12847/) (2020)
- [The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, Subjective Testing Framework, and Challenge Results](https://mlsanity.com/paper/2005.13981/) (2020)
- [MetricGAN: Generative Adversarial Networks based Black-box Metric Scores Optimization for Speech Enhancement](https://mlsanity.com/paper/1905.04874/) (2019)
- [MetricGAN+: An Improved Version of MetricGAN for Speech Enhancement](https://mlsanity.com/paper/2104.03538/) (2021)
- [FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement](https://mlsanity.com/paper/2010.15508/) (2020)
- [A Hybrid DSP/Deep Learning Approach to Real-Time Full-Band Speech Enhancement](https://mlsanity.com/paper/1709.08243/) (2017)
- [TSTNN: Two-stage Transformer based Neural Network for Speech Enhancement in the Time Domain](https://mlsanity.com/paper/2103.09963/) (2021)
- [TinyLSTMs: Efficient Neural Speech Enhancement for Hearing Aids](https://mlsanity.com/paper/2005.11138/) (2020)
- [Replay and Synthetic Speech Detection with Res2net Architecture](https://mlsanity.com/paper/2010.15006/) (2020)
- [PoCoNet: Better Speech Enhancement with Frequency-Positional Embeddings, Semi-Supervised Conversational Data, and Biased Loss](https://mlsanity.com/paper/2008.04470/) (2020)
- [Improved Speech Enhancement with the Wave-U-Net](https://mlsanity.com/paper/1811.11307/) (2018)
- [Speech Denoising in the Waveform Domain with Self-Attention](https://mlsanity.com/paper/2202.07790/) (2022)
- [WaveCRN: An Efficient Convolutional Recurrent Neural Network for End-to-end Speech Enhancement](https://mlsanity.com/paper/2004.04098/) (2020)
- [Investigating Cross-Domain Losses for Speech Enhancement](https://mlsanity.com/paper/2010.10468/) (2020)
- [Noisy Training Improves E2E ASR for the Edge](https://mlsanity.com/paper/2107.04677/) (2021)

## Built on by (most cited first, 1 of 1)

- [Reverse Attention for Lightweight Speech Enhancement on Edge Devices](https://mlsanity.com/paper/2509.16705/) (2025)

## Abstract (arXiv)

Deep learning has become a de facto method of choice for speech enhancement tasks with significant improvements in speech quality. However, real-time processing with reduced size and computations for low-power edge devices drastically degrades speech quality. Recently, transformer-based architectures have greatly reduced the memory requirements and provided ways to improve the model performance through local and global contexts. However, the transformer operations remain computationally heavy. In this work, we introduce WaveUNet squeeze-excitation Res2 (WSR)-based metric generative adversarial network (WSR-MGAN) architecture that can be efficiently implemented on low-power edge devices for noise suppression tasks while maintaining speech quality. We utilize multi-scale features using Res2Net blocks that can be related to spectral content used in speech-processing tasks. In the generator, we integrate squeeze-excitation blocks (SEB) with multi-scale features for maintaining local and global contexts along with gated recurrent units (GRUs). The proposed approach is optimized through a combined loss function calculated over raw waveform, multi-resolution magnitude spectrogram, and objective metrics using a metric discriminator. Experimental results in terms of various objective metrics on VoiceBank+DEMAND and DNS-2020 challenge datasets demonstrate that the proposed speech enhancement (SE) approach outperforms the baselines and achieves state-of-the-art (SOTA) performance in the time domain.