Active contour method for ILM segmentation in ONH volume scans in retinal OCT

Area of Science:

• Ophthalmology
• Medical Imaging
• Computational Biology

Background:

• The optic nerve head (ONH) is crucial for vision and affected by various diseases.
• Accurate segmentation of ONH structures is vital for diagnosis and monitoring.
• Current methods struggle with the ONH's complex anatomy and image artifacts.

Purpose of the Study:

• To develop a robust method for segmenting the inner limiting membrane (ILM) in 3D ONH scans.
• To address challenges posed by complex anatomy and inhomogeneous image intensities.
• To improve the accuracy of ILM segmentation compared to existing solutions.

Main Methods:

• Utilized a Chan-Vese active contour model with local intensity fitting energy.
• Incorporated a boundary potential to prevent segmentation of outer retinal layers.
• Employed local rescaling of intensities and differential evolution for parameter optimization.

Main Results:

• The proposed method demonstrated robust ILM segmentation in challenging ONH scans.
• Successfully handled inhomogeneous image intensities and complex topological structures.
• Significantly outperformed a commercial segmentation solution in evaluations.

Conclusions:

• The developed active contour method offers a significant advancement in ONH image segmentation.
• This technique has the potential to improve the diagnosis and management of ONH diseases.
• Further research can explore its application in other complex biological imaging tasks.