Human Swallowing Project

Benjamin D'incau

Semester Project

Section of Communication Systems, EPFL

January 2013

Abstract

The purpose of the project was to provide a new computational technique to help doctors to diagnose patients with swallowing disorders through the development of a unique application in image processing.
Human Swallowing is a vital function for the human being. A proper ingestion of food and liquid is critical to survival and symptoms associated with swallowing impairment (or Dysphagia) can be dramatic, and sometimes delayed.
The hyoid bone (located beneath the jaw) maximum displacement from the vertebra is the first clinical value that enables doctors and medecine specialists to diagnose dysphagic patients.
This plugin helps with this diagnosis by finding the maximum hyoid displacement within a .TIFF file, previsouly recorded via Video Fluoroscopy (X-Ray). Using mainly Template Matching Criterions (Sum of Absolute Differences, Optimized Sum of Absolute Differences and Mean Squared Error), the plugin can output the Maximum Hyoid Displacement value in milimeters, as well as the location of the bolus (food or liquid intake) at that time.