ASHA journals
Browse
1/1
10 files

Perspective distortion in laryngoscopy (Veltrup et al., 2023)

online resource
posted on 2023-08-31, 22:43 authored by Reinhard Veltrup, Stefan Kniesburges, Marion Semmler

Objective: An experiment with controllable boundaries was designed to assess the influence of the recording angle and distance on two-dimensional (2D) imaging in laryngoscopy and resulting 2D parameter calculation derived from the glottal area waveform (GAW).

Method: Two high-speed camera setups were used to synchronously record an oscillating synthetic vocal fold (VF) model, simulating a high-speed videoendoscopy. One camera recorded at variable lateral recording angles and a reference camera in superior perspective. This was performed at different physiological recording distances and for two oscillation modes (with/without contacting VFs). The GAW was derived from the segmented glottis, and two parameters each for the categories of symmetry, periodicity, and closure were calculated, as well as two derivative measures. The percentage difference between the variable and reference camera value pairs was calculated, and the angle and height dependencies were quantified using linear regression.

Results: The visual perception of a laryngoscopy was found to be influenced by the lateral recording angle, which may lead to misinterpretation of VF symmetry among inexperienced observers. The strongest influence of recording angle was observed for symmetry parameters, the strongest being the Amplitude Symmetry Index with up to 2.6%/° (p < .05). A dependence on the recording distance was only found for the Maximum Area Declination Rate.

Conclusions: The recording angle in 2D laryngoscopy should be carefully considered during visual inspection of the VF dynamics. Most of the investigated objective parameters were unaffected by the examined perspective distortion. However, especially left–right symmetry measures should only be used under controlled boundary conditions to avoid misdiagnosis and misinterpretation.

Supplemental Material S1. Linear regression results of the parameters over the recording angles α from the variable camera (m: slope [%/°], b: constant [%], p: probability that the relationship between the angle and the parameter change is random).

Supplemental Material S2. Linear regression results of the parameter change over the three camera distances (50mm/65mm/80mm) from the variable camera (m: slope [%/mm], b: constant [%], p: probability that the relationship between the distance and the parameter change is random).

Supplemental Material S3. Original video of the reference camera at a recording angle of 0° and oscillation mode M1. The recording frequency is 4kHz.

Supplemental Material S4. Original video of the reference camera at a recording angle of 0° and oscillation mode M2. The recording frequency is 4kHz.

Supplemental Material S5. Original video of the variable camera at a recording angle of 0° the oscillation mode M1. The recording frequency is 4kHz.

Supplemental Material S6. Original video of the variable camera at a recording angle of 0° and oscillation mode M2. The recording frequency is 4kHz.

Supplemental Material S7. Original video of the variable camera at a recording angle of 5° and oscillation mode M1. The recording frequency is 4kHz.

Supplemental Material S8. Original video of the variable camera at a recording angle of 5° and oscillation mode M2. The recording frequency is 4kHz.

Supplemental Material S9. Original video of the variable camera at a recording angle of 10° and oscillation mode M1. The recording frequency is 4kHz.

Supplemental Material S10. Original video of the variable camera at a recording angle of 10° and oscillation mode M2. The recording frequency is 4kHz.

Veltrup, R., Kniesburges, S., & Semmler, M. (2023). Influence of perspective distortion in laryngoscopy. Journal of Speech, Language, and Hearing Research, 66(9), 3276–3289. https://doi.org/10.1044/2023_JSLHR-23-00027

Funding

We acknowledge the support by Deutsche Forschungsgemeinschaft Grant No. DO 1247/16-1 (awarded to Michael Döllinger) under Project Number 448240908.

History