A video-based speed estimation technique for localizing the wireless capsule endoscope inside gastrointestinal tract

Area of Science:

• Medical Imaging
• Computer Vision
• Gastroenterology

Background:

• Wireless Capsule Endoscopy (WCE) is a key non-invasive tool for gastrointestinal (GI) tract inspection.
• Accurate localization of the endoscopic capsule is crucial for identifying intestinal disease positions.
• Precise measurement of capsule travel distance requires accurate speed estimation.

Purpose of the Study:

• To present a novel computer vision-based speed estimation technique for WCE.
• To enhance the localization accuracy of WCE through precise speed measurement.

Main Methods:

• Analyzing displacements between consecutive frames from WCE videos.
• Developing a computer vision algorithm to extract capsule speed.
• Validating the approach using both virtual testbeds and real endoscopic images.

Main Results:

• The proposed method achieves an average accuracy of 93% in estimating endoscopic capsule speed.
• The enhanced localization accuracy of WCE was improved to less than 2.49 cm.
• The technique effectively extracts capsule speed from image data.

Conclusions:

• The developed computer vision technique accurately estimates WCE capsule speed.
• This speed estimation significantly improves the precision of WCE localization.
• The method offers a valuable tool for identifying intestinal disease positions during capsule endoscopy.