Automatic anterior chamber angle structure segmentation in AS-OCT image based on label transfer.

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Vision

Background:

• The anterior chamber angle (ACA) is critical for diagnosing and treating angle-closure glaucoma.
• Anterior Segment Optical Coherence Tomography (AS-OCT) provides structural assessment of the ACA.
• Accurate segmentation of ACA structures is essential for clinical applications.

Purpose of the Study:

• To develop a novel, fully automatic method for segmenting anterior chamber angle structures in AS-OCT images.
• To extract key clinical structures including the corneal boundary, iris region, and trabecular-iris contact.
• To validate the method's performance on a clinical AS-OCT dataset.

Main Methods:

• Utilizing label transfer from a reference AS-OCT dataset to obtain initial labels.
• Refining initial labels to serve as landmarks for supporting structure segmentation.
• Implementing a segmentation process to extract corneal boundary, iris region, and trabecular-iris contact.

Main Results:

• The proposed method achieved satisfactory segmentation performance on clinical AS-OCT data.
• Successfully extracted major clinical structures of the anterior chamber angle.
• Demonstrated the feasibility of automatic segmentation for ACA analysis.

Conclusions:

• The novel automatic segmentation method shows significant potential for clinical ACA parameter measurement.
• This approach can contribute to the development of automatic glaucoma classification systems.
• The technique offers a promising tool for enhancing glaucoma diagnosis and treatment strategies.