Summary
Objectives: The most radical cancer therapy of the throat is the total excision of the larynx
which postoperatively results in the loss of voice. A widely-used method of voice
rehabilitation is the insertion of a silicone valve, which establishes an unidirectional
connection between trachea and esophagus. Thus, during exhalation, air can be directed
from the trachea into the esophagus. This air stream excites tissue vibrations of
the esophagus and the hypo-pharynx which act as a substitute voice generator. Purpose
of the current study is to present a technique for visualizing the dynamics of the
substitute voice generating element.
Methods: Digital high speed videos of the vibrating tissue are simultaneously recorded with
the emitted acoustic signal. The high speed sequences are directly evaluated by a
three-step knowledge based algorithm. It considers correlation between image and acoustic
data, information about the gray value of each pixel, and continuity of tissue vibration.
The temporal properties of an image series are investigated by evaluating the time
dependent gray value at each pixel position.
Results: The applicability of the algorithm is exemplar-ily demonstrated using the data of
one male patient. It enables the identification of the regions within an image series
which are mainly responsible for the acoustic signal. Additionally, the dynamics of
tissue vibrations are visualized. The main propagation direction can be clearly identified.
Conclusions: The new methodology summarizes the information about endoscopic and acoustic recordings
of substitute voice into a single image. The results allow a first estimation of tissue
velocity and elastic properties of oscillating tissue.
Keywords
Substitute voice - image processing - digital high speed camera - PE-segment
