Thickness dependent tortuosity estimation for retinal blood vessels

Area of Science:

• Ophthalmology
• Medical Imaging
• Computational Biology

Background:

• Vessel tortuosity in retinal images is a key indicator of various vascular diseases.
• Accurate and automated measurement of retinal vessel tortuosity is crucial for diagnosis and monitoring.
• Existing tortuosity metrics may not fully capture the complexity of retinal vasculature.

Purpose of the Study:

• To develop and validate a novel framework for automated estimation of vessel tortuosity in retinal images.
• To introduce a new tortuosity metric that incorporates vessel thickness for improved accuracy.
• To compare the proposed method's performance against clinical judgment and existing metrics.

Main Methods:

• A framework for automated retinal vessel tortuosity estimation was developed.
• A new tortuosity metric considering vessel thickness was introduced.
• A graph-based algorithm utilizing medial axis representation and shortest-path analysis was employed for segment identification and tortuosity calculation.

Main Results:

• The proposed tortuosity metric demonstrated estimates closer to medical intuition than previous metrics.
• The automated framework achieved an overall agreement of 92.4% with clinical judgment on 50 retinal vessels from the DRIVE dataset.
• The developed method outperformed comparison measures in accuracy.

Conclusions:

• The proposed framework offers a robust and accurate method for automated retinal vessel tortuosity estimation.
• Incorporating vessel thickness into tortuosity metrics enhances their clinical relevance and agreement with expert assessment.
• This automated approach has the potential to aid in the early detection and management of retinal vascular diseases.