SLPOtherJournal of voice : official journal of the Voice Foundation2023

Spatial Segmentation for Laryngeal High-Speed Videoendoscopy in Connected Speech.

Ahmed M Yousef, Dimitar D Deliyski, Stephanie R C Zacharias and 3 others

PMID 33257208

WHAT IT FOUND

An automated method tracked vocal fold edges in one healthy speaker reading aloud.

It worked for 67 of 76 vocalizations, failing only when lighting was poor. This shows the algorithm is feasible but not yet proven for patients with voice disorders.

Key findings

01The algorithm successfully detected vocal fold edges in 67 of 76 vocalizations from a single healthy speaker, with an error of no more than one pixel.

02The method failed to detect edges in 9 vocalizations due to dim lighting in the high-speed video frames.

03The study was limited to one vocally normal participant, so generalization to patients with voice disorders requires further research.

STILL TO COME

How it was doneWhat they foundWhat it means for SLPs

Read the rest of this summary

You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.

Already have one?

What it does not show

The study involved only one vocally normal participant, so the algorithm's performance in patients with voice disorders is unknown. The method failed in 9 of 76 vocalizations due to poor lighting, indicating sensitivity to image quality. The algorithm struggled to detect edges during non-vibratory phases (before onset and after offset of phonation). The results are based on visual inspection of a single recording, not a large-scale validation study. The study used color images, which are more challenging for segmentation than monochrome images, but the authors claim monochrome results would be better without showing them.

Declared interests

The authors declare no conflict of interest.

The easy way to misread this

Do not interpret this as evidence that automated HSV analysis is ready for clinical use. The algorithm was tested on a single healthy speaker and failed for 12% of vocalizations due to lighting issues. It has not been validated on patients with voice disorders.

Read it on PubMed →