Effects of Deep Learning Architecture and Pretraining on Automated Vocal Fold Segmentation in Ex Vivo High-Speed Phonation Images
Abstract
Two-dimensional (2D) intraglottal particle image velocimetry (PIV) of ex vivo phonation experiments quantifies vocal fold medial surface dynamics and the intraglottal flow field. However, it relies on manual segmentation because residual glare, low tissue-to-flow contrast, and tracer-particle speckle obscure the vocal fold boundaries. This study assessed how deep learning architecture and ImageNet pretraining affect automated vocal fold segmentation in these high-speed images. Four established architectures from three categories were benchmarked: the convolutional neural network (CNN)-based UNet and UNet++, the hybrid CNN-transformer-based TransUNet, and the pure transformer-based Swin-UNet, the latter two trained from scratch and from ImageNet-pretrained weights. Using 16,344 expert-annotated images from 227 phonation trials of excised canine larynges, the six pipelines were trained with five seeds under a standardized protocol and compared on 50 held-out test trials using trial-level Wilcoxon signed-rank tests with Holm correction. Mean Dice similarity coefficient (DSC) spanned only 0.9734 to 0.9764, so the pipelines differed mainly in boundary fidelity. Pretraining significantly improved the boundary F1 score (BF1) and 95th percentile Hausdorff distance (HD95) of the TransUNet and Swin-UNet, about four times more for the Swin-UNet. Once pretrained, the two performed similarly, although the Swin-UNet has only about a quarter of the trainable parameters. Without pretraining, the TransUNet had a significantly higher BF1 and a lower HD95 than both CNN-based pipelines and about half their share of images with HD95 > 50 pixels. To our knowledge, this is the first study to apply deep learning to vocal fold segmentation in 2D intraglottal PIV, and it offers a reference framework for segmenting experimental flow-visualization images with noisy optical conditions or highly deformable objects.
Related
- SuperJEPA: Pretraining with a Supervision-Augmented Joint-Embedding Predictive Architecture for High Energy Physics
- Automated carotid artery atherosclerotic inflammation segmentation in PET using MSR-Net
- Robust acoustic representation of Japanese vowels under phonation variability using dynamic spectral analysis
- Post-Selection-Free Quantum Automated Learning
- Improving Clinical Target Volume Segmentation Accuracy using Anatomical Priors and Active Learning for the AGITG TOPGEAR Clinical Trial
BibTeX
@misc{kalaiarasan2026effects,
title = {{Effects of Deep Learning Architecture and Pretraining on Automated Vocal Fold Segmentation in Ex Vivo High-Speed Phonation Images}},
author = {Varun Vinayak Kalaiarasan and Charles Farbos de Luzan and Khalid Malik and Ephraim Gutmark and Liran Oren},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.37860},
eprint = {2610.37860},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.37860},
}
RIS
TY - UNPB
TI - Effects of Deep Learning Architecture and Pretraining on Automated Vocal Fold Segmentation in Ex Vivo High-Speed Phonation Images
AU - Varun Vinayak Kalaiarasan
AU - Charles Farbos de Luzan
AU - Khalid Malik
AU - Ephraim Gutmark
AU - Liran Oren
PY - 2026
DA - 2026/10/27
PB - arXiv
DO - 10.48550/arxiv.2610.37860
UR - https://arxiv.org/abs/2610.37860
AB - Two-dimensional (2D) intraglottal particle image velocimetry (PIV) of ex vivo phonation experiments quantifies vocal fold medial surface dynamics and the intraglottal flow field. However, it relies on manual segmentation because residual glare, low tissue-to-flow contrast, and tracer-particle speckle obscure the vocal fold boundaries. This study assessed how deep learning architecture and ImageNet pretraining affect automated vocal fold segmentation in these high-speed images. Four established architectures from three categories were benchmarked: the convolutional neural network (CNN)-based UNet and UNet++, the hybrid CNN-transformer-based TransUNet, and the pure transformer-based Swin-UNet, the latter two trained from scratch and from ImageNet-pretrained weights. Using 16,344 expert-annotated images from 227 phonation trials of excised canine larynges, the six pipelines were trained with five seeds under a standardized protocol and compared on 50 held-out test trials using trial-level Wilcoxon signed-rank tests with Holm correction. Mean Dice similarity coefficient (DSC) spanned only 0.9734 to 0.9764, so the pipelines differed mainly in boundary fidelity. Pretraining significantly improved the boundary F1 score (BF1) and 95th percentile Hausdorff distance (HD95) of the TransUNet and Swin-UNet, about four times more for the Swin-UNet. Once pretrained, the two performed similarly, although the Swin-UNet has only about a quarter of the trainable parameters. Without pretraining, the TransUNet had a significantly higher BF1 and a lower HD95 than both CNN-based pipelines and about half their share of images with HD95 > 50 pixels. To our knowledge, this is the first study to apply deep learning to vocal fold segmentation in 2D intraglottal PIV, and it offers a reference framework for segmenting experimental flow-visualization images with noisy optical conditions or highly deformable objects.
ER -