A torsional eye movement calculation algorithm for low contrast images in video-oculography

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Computer Vision

Background:

• Video-oculography (VOG) is crucial for detecting eye movements.
• Existing algorithms struggle with torsional eye movement detection, especially in low-contrast images common in certain clinical tests.
• Low image contrast due to IR-illumination makes iris textures invisible, hindering standard torsional movement algorithms.

Purpose of the Study:

• To design and implement a robust algorithm for detecting torsional eye movements in VOG.
• To address the challenge of measuring torsional movements in low-contrast video images.
• To provide a more reliable method for VOG analysis in specific clinical scenarios.

Main Methods:

• Development of a novel algorithm for torsional eye movement detection.
• Utilizing head-mounted video cameras and IR-illumination for eye imaging.
• Implementation of the algorithm in a clinical device for real-world testing.

Main Results:

• The developed algorithm demonstrates robust performance in detecting torsional eye movements.
• The new approach is effective even with very low image contrast.
• Performance comparison with alternative techniques was conducted.

Conclusions:

• The novel algorithm offers a significant improvement for torsional eye movement detection in VOG.
• This method enhances VOG's utility in clinical settings requiring low-contrast imaging.
• The implemented algorithm provides a viable solution for challenging VOG analysis.