Intra-retinal layer segmentation in optical coherence tomography using an active contour approach.

Area of Science:

• Ophthalmology
• Biomedical Imaging
• Computational Biology

Background:

• Optical coherence tomography (OCT) is a key non-invasive imaging technique in ophthalmology.
• Retinal degeneration models in rodents are crucial for studying eye diseases.
• Accurate segmentation of intra-retinal layers is challenging due to low contrast and noise in OCT images.

Purpose of the Study:

• To develop an automatic segmentation algorithm for intra-retinal layers in rodent OCT images.
• To adapt and improve active contour models for noisy and low-contrast OCT data.
• To validate the algorithm's accuracy, especially for the ganglion cell layer relevant to glaucoma.

Main Methods:

• Adaptation of Chan-Vese's energy-minimizing active contours without edges.
• Implementation of a multi-phase framework with a circular shape prior.
• Utilizing a contextual scheme to balance energy functional terms and least squares for parameter estimation.

Main Results:

• The algorithm demonstrated strong performance in segmenting intra-retinal layers in OCT images.
• Achieved an average Dice similarity coefficient of 0.85 across 20 images from four rats.
• Specifically, achieved a Dice similarity coefficient of 0.94 for the ganglion cell layer.

Conclusions:

• The developed automatic segmentation algorithm is effective for detecting intra-retinal layers in rodent OCT images.
• The method shows high accuracy, particularly for the ganglion cell layer, with potential applications in glaucoma research.
• The algorithm's robustness to noise and low contrast makes it suitable for challenging OCT datasets.