Fundus image mosaicking for information augmentation in computer-assisted slit-lamp imaging

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Vision

Background:

• Laser photocoagulation is standard for diabetic retinopathy and retinal vein occlusions.
• Slit-lamp biomicroscopes offer limited visualization of the retina, complicating treatment.
• Current methods hinder target identification and increase procedure time.

Purpose of the Study:

• To develop an intraoperative retina mapping system for view expansion using slit-lamp devices.
• To enhance surgical navigation and efficiency in treating retinal diseases.
• To improve visualization for photo documentation and diagnosis.

Main Methods:

• A mosaicking method combining direct and feature-based approaches was adapted.
• Enhancements include visual tracking with illumination compensation.
• An efficient pixel selection and entropy-based mosaic update were implemented.

Main Results:

• The system demonstrated improved visualization and navigation capabilities in human subject experiments.
• The method effectively handles challenging visualization conditions.
• Enhanced computational efficiency was achieved.

Conclusions:

• The proposed retina mapping system offers practical value for intraoperative navigation, photo documentation, and diagnosis.
• This technology can assist surgeons in treating sight-threatening retinal diseases more effectively.
• Further development can enhance surgical outcomes and patient comfort.