# Reconstructing Speech from Real-Time Articulatory MRI Using Neural Vocoders

arXiv: https://arxiv.org/abs/2104.11598 · HTML: https://mlsanity.com/paper/2104.11598/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Apr 2021.

## Built on by (most cited first, 6 of 6)

- [Multimodal Segmentation for Vocal Tract Modeling](https://mlsanity.com/paper/2406.15754/) (2024)
- [MRI2Speech: Speech Synthesis from Articulatory Movements Recorded by Real-time MRI](https://mlsanity.com/paper/2412.18836/) (2024)
- [Tagged-MRI Sequence to Audio Synthesis via Self Residual Attention Guided Heterogeneous Translator](https://mlsanity.com/paper/2206.02284/) (2022)
- [Adaptation of Tongue Ultrasound-Based Silent Speech Interfaces Using Spatial Transformer Networks](https://mlsanity.com/paper/2305.19130/) (2023)
- [Automated Identification of Failure Cases in Organ at Risk Segmentation Using Distance Metrics: A Study on CT Data](https://mlsanity.com/paper/2308.10636/) (2023)
- [Multimodal Deep Learning for Subtype Classification in Breast Cancer Using Histopathological Images and Gene Expression Data](https://mlsanity.com/paper/2503.02849/) (2025)

## Abstract (arXiv)

Several approaches exist for the recording of articulatory movements, such as eletromagnetic and permanent magnetic articulagraphy, ultrasound tongue imaging and surface electromyography. Although magnetic resonance imaging (MRI) is more costly than the above approaches, the recent developments in this area now allow the recording of real-time MRI videos of the articulators with an acceptable resolution. Here, we experiment with the reconstruction of the speech signal from a real-time MRI recording using deep neural networks. Instead of estimating speech directly, our networks are trained to output a spectral vector, from which we reconstruct the speech signal using the WaveGlow neural vocoder. We compare the performance of three deep neural architectures for the estimation task, combining convolutional (CNN) and recurrence-based (LSTM) neural layers. Besides the mean absolute error (MAE) of our networks, we also evaluate our models by comparing the speech signals obtained using several objective speech quality metrics like the mean cepstral distortion (MCD), Short-Time Objective Intelligibility (STOI), Perceptual Evaluation of Speech Quality (PESQ) and Signal-to-Distortion Ratio (SDR). The results indicate that our approach can successfully reconstruct the gross spectral shape, but more improvements are needed to reproduce the fine spectral details.