Department of Image Processing

We have proposed methodology for the enhancement of videokymographic data and the extraction of typical characteristics of vocal folds vibrations, such as, e.g., the time-varying extent of rima glottidis and the progression of mucosal waves, and corresponding vibration parameters (e.g. frequency, symmetries, amplitude). It is based on well-established digital image processing methods; namely image denoising, edge detection, image segmentation, and object identification. The set of evaluated features was designed by experts, drawing on their longtime expertise with videokymography. These features are mostly based on the estimated shape of the glottal contour, i.e. the boundary between the vocal folds and the rima glottidis, but also on mucosal waves.

An example of the VKG features and parameters

Publications: