Automatic segmentation of choroidal thickness in optical coherence tomography.

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• Choroidal thickness assessment from optical coherence tomography (OCT) images is crucial for understanding eye health and disease.
• Manual segmentation of choroidal boundaries is time-consuming and subjective, necessitating automated methods.
• Accurate choroidal thickness measurement is vital for diagnosing eye diseases and refractive errors.

Purpose of the Study:

• To develop and validate an automated graph-search based segmentation technique for accurate choroidal thickness measurement from OCT images.
• To address the challenges of non-uniform tissue, low contrast, and artifacts in choroidal boundary detection.
• To provide a reliable objective method for deriving choroidal thickness profiles.

Main Methods:

• An automated segmentation technique utilizing graph-search theory was developed.
• Pre-processing steps were applied to enhance boundaries and minimize artifacts in OCT B-scans.
• Specific algorithms were designed for detecting the inner choroidal boundary (ICB) and outer choroidal boundary (OCB) using edge filters, weighted maps, and probability gradients.

Main Results:

• The proposed method demonstrated robust detection of both ICB and OCB in OCT images.
• Validation on large pediatric and adult datasets showed high accuracy compared to manual segmentation.
• The automated technique proved effective in extracting clinically relevant choroidal thickness data.

Conclusions:

• The developed graph-search based method offers an objective and efficient approach for choroidal segmentation in OCT images.
• This automated technique is a valuable tool for clinical data extraction and advancing eye research.
• Accurate choroidal thickness measurement using this method can aid in the diagnosis and management of various eye conditions.