A statistical shape space model of the palate surface trained on 3D MRI scans of the vocal tract

TL;DR

This paper introduces a minimally-supervised statistical shape model of the palate surface derived from 3D MRI scans, demonstrating low error and broad applicability to various imaging modalities.

Contribution

A novel PCA-based method for creating a statistical palate shape model from MRI data with minimal supervision and cross-modality applicability.

Findings

Low reconstruction error with limited measured coordinates

Model generalizes well to new MRI and optical scan data

Applicable to other modalities like EMA and UTI

Abstract

We describe a minimally-supervised method for computing a statistical shape space model of the palate surface. The model is created from a corpus of volumetric magnetic resonance imaging (MRI) scans collected from 12 speakers. We extract a 3D mesh of the palate from each speaker, then train the model using principal component analysis (PCA). The palate model is then tested using 3D MRI from another corpus and evaluated using a high-resolution optical scan. We find that the error is low even when only a handful of measured coordinates are available. In both cases, our approach yields promising results. It can be applied to extract the palate shape from MRI data, and could be useful to other analysis modalities, such as electromagnetic articulography (EMA) and ultrasound tongue imaging (UTI).