Retinal pigment epithelium segmentation by polarization sensitive optical coherence tomography

Area of Science:

• Ophthalmology
• Biomedical Imaging
• Optics

Background:

• Accurate segmentation of the retinal pigment epithelium (RPE) is crucial for diagnosing retinal diseases.
• Existing intensity-based segmentation methods for RPE in optical coherence tomography (OCT) images have limitations.
• Polarization sensitive OCT (PS-OCT) offers unique tissue property information beyond intensity.

Purpose of the Study:

• To develop and present a novel method for identifying and segmenting the RPE layer in human retinal images.
• To leverage the intrinsic depolarizing property of the RPE for improved segmentation accuracy.
• To demonstrate the efficacy of the proposed method in both healthy and diseased retinas using PS-OCT.

Main Methods:

• Developed two segmentation algorithms utilizing polarization properties of the RPE, specifically its light depolarizing effect.
• Algorithm 1: Based on retardation data derived from PS-OCT.
• Algorithm 2: A more advanced method using local polarization state variations calculated from averaged Stokes vector elements.
• Utilized a state-of-the-art spectral domain PS-OCT instrument for data acquisition.

Main Results:

• Successfully demonstrated a new approach for RPE segmentation based on intrinsic tissue depolarization.
• Presented two distinct algorithms, one simpler (retardation-based) and one more complex (Stokes vector-based).
• Validated the method's performance in identifying RPE in both healthy and diseased human retinal PS-OCT images.

Conclusions:

• The proposed polarization-based segmentation method offers a novel alternative to traditional intensity-based approaches for RPE identification.
• The intrinsic depolarizing property of the RPE is a valuable feature for accurate segmentation in PS-OCT imaging.
• This technique shows promise for enhanced diagnostic capabilities in retinal imaging of various ocular conditions.