# Improving Visual Speech Enhancement Network by Learning Audio-visual Affinity with Multi-head Attention

arXiv: https://arxiv.org/abs/2206.14964 · HTML: https://mlsanity.com/paper/2206.14964/

## Key facts

- Cited by 4 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 6 of its 30 references that are in the map.
- First posted on arXiv 30 Jun 2022.

## 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, 6 of 6)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)](https://mlsanity.com/paper/1511.07289/) (2015)
- [Audio-Visual Speech Enhancement Using Multimodal Deep Convolutional Neural Networks](https://mlsanity.com/paper/1709.00944/) (2017)
- [WaveCRN: An Efficient Convolutional Recurrent Neural Network for End-to-end Speech Enhancement](https://mlsanity.com/paper/2004.04098/) (2020)
- [Deep-Learning-Based Audio-Visual Speech Enhancement in Presence of Lombard Effect](https://mlsanity.com/paper/1905.12605/) (2019)

## Built on by (most cited first, 4 of 4)

- [Real-Time Audio-Visual End-to-End Speech Enhancement](https://mlsanity.com/paper/2303.07005/) (2023)
- [Incorporating Ultrasound Tongue Images for Audio-Visual Speech Enhancement through Knowledge Distillation](https://mlsanity.com/paper/2305.14933/) (2023)
- [Incorporating Ultrasound Tongue Images for Audio-Visual Speech Enhancement](https://mlsanity.com/paper/2309.10455/) (2023)
- [Enhancing Neural Speech Coding with Semantic and Visual Cues](https://mlsanity.com/paper/2609.05076/) (2026)

## Abstract (arXiv)

Audio-visual speech enhancement system is regarded as one of promising solutions for isolating and enhancing speech of desired speaker. Typical methods focus on predicting clean speech spectrum via a naive convolution neural network based encoder-decoder architecture, and these methods a) are not adequate to use data fully, b) are unable to effectively balance audio-visual features. The proposed model alleviates these drawbacks by a) applying a model that fuses audio and visual features layer by layer in encoding phase, and that feeds fused audio-visual features to each corresponding decoder layer, and more importantly, b) introducing a 2-stage multi-head cross attention (MHCA) mechanism to infer audio-visual speech enhancement for balancing the fused audio-visual features and eliminating irrelevant features. This paper proposes attentional audio-visual multi-layer feature fusion model, in which MHCA units are applied to feature mapping at every layer of decoder. The proposed model demonstrates the superior performance of the network against the state-of-the-art models.