A preliminary study of a quantitative analysis method for high speed laryngoscopic images.

ABSTRACT

High speed laryngoscopic study provides new information on the vocal fold vibratory patterns which has not been made possible before. Qualitative analysis of high speed laryngoscopic images is commonly reported in the literature. However, quantitative studies are necessary to provide objective measures for research and clinical purposes. This study reports the development of the High Speed Video Processing Program in analysing high-speed laryngoscopic video images obtained from 15 non-dysphonic speakers. Nine glottal ratio indices that represented the full opened glottal area, glottal width and glottal length were computed using the processing program. Data from two dysphonic participants were included to illustrate the manifestation of these glottal measures in dysphonic voices. Some of these glottal ratio index measures showed high variability across participants. However, significant differences were found in a number of index measures among the creaky, modal and falsetto registers in the non-dysphonic participants. In addition, the glottal area ratio index was found to be specifically sensitive in detecting dysphonia in high speed glottal images. The proposed semi-automatic High Speed Video Processing Program showed value in the analysis of high speed laryngoscopic images.