Semi-automatic assessment of hyoid bone motion in digital videofluoroscopic images

Ishtiaque Hossain, Angela Roberts-South, Mandar Jog, Mahmoud R. El-Sakka

Research output: Contribution to journalArticlepeer-review

9 Scopus citations


The swallowing process involves triggering the movements of a number of muscles in the throat that transports the food from the mouth to the stomach successfully and at the same time prevents it from getting into the airway and the lung. In order to detect abnormalities in the swallowing process, radiologists use a technique called videofluoroscopic swallowing study. It is a video of X-ray images that are taken while the patient swallows food, which is later visually inspected by the radiologist to evaluate the patient's swallowing ability. It has been reported that measuring the movement of the hyoid bone plays an important role in the evaluation process. However, due to the subjective nature of visual inspection, radiologists have difficulty reaching unanimous decision about the outcome of the evaluation. In this research, a semi-automatic method is proposed which tracks the hyoid bone and quantifies its movement. Using a classification-based approach, the proposed method automatically identifies the region of interest before identifying the hyoid bone. This allows limiting image processing procedures to the relevant area in the image. Results show that the proposed method identifies and tracks the hyoid bone with significant accuracy.

Original languageEnglish (US)
Pages (from-to)25-37
Number of pages13
JournalComputer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization
Issue number1
StatePublished - Jan 2014


  • Haar classifier
  • dysphagia
  • hyoid bone
  • object detection
  • tracking
  • videofluoroscopic swallowing study

ASJC Scopus subject areas

  • Computational Mechanics
  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging
  • Computer Science Applications


Dive into the research topics of 'Semi-automatic assessment of hyoid bone motion in digital videofluoroscopic images'. Together they form a unique fingerprint.

Cite this