Low rank based FAZ segmentation in OCTA images.

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• Optical Coherence Tomography Angiography (OCTA) is a key non-invasive technique for retinal vascular imaging.
• Retinal diseases can alter vascular density and the Fovea Avascular Zone (FAZ) morphology.
• FAZ analysis is crucial for diagnosing various eye conditions.

Purpose of the Study:

• To develop and evaluate novel methods for accurate FAZ segmentation in OCTA images.
• To improve the diagnosis of eye diseases through precise FAZ analysis.
• To introduce low-rank Tensor Ring (TR) decomposition for FAZ segmentation.

Main Methods:

• Proposed a three-step approach for FAZ segmentation: localization, de-noising, and refinement.
• Employed two FAZ localization techniques: Low-rank Tensor Ring (TR) decomposition and morphological operators.
• Utilized de-noised OCTA images and FAZ localization information for segmentation.

Main Results:

• Achieved accurate segmentation of the Fovea Avascular Zone (FAZ) from OCTA images.
• Demonstrated the effectiveness of the proposed methods, including the novel use of low-rank TR estimation.
• Simulation results validated the performance of the segmentation techniques.

Conclusions:

• The proposed methods offer accurate FAZ segmentation from OCTA images.
• Low-rank TR decomposition is a novel and effective approach for FAZ segmentation.
• Accurate FAZ segmentation can significantly aid in the early diagnosis of retinal diseases.