Automated layer segmentation of optical coherence tomography images

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer-Aided Diagnosis

Background:

• Optical coherence tomography (OCT) is crucial for measuring retinal nerve fiber layer thickness.
• Accurate retinal layer segmentation is essential for glaucoma diagnosis using OCT.
• Existing methods may struggle with noise and preserving global image variations.

Purpose of the Study:

• To present an automated retinal layer segmentation technique for OCT images.
• To enhance the accuracy of computer-aided diagnosis in ophthalmology.
• To develop a robust method for identifying retinal layers in OCT scans.

Main Methods:

• OCT images are segmented into vessel and nonvessel sections using iterative polynomial smoothing for blood vessel detection.
• Nonvessel sections undergo bilateral and median filtering to reduce noise while preserving layer boundaries.
• Retinal layer boundaries are detected and classified from filtered nonvessel sections.

Main Results:

• The proposed technique accurately segments OCT images into five distinct retinal layers.
• The method effectively suppresses local noise while maintaining global image variations.
• Experimental results demonstrate high accuracy in retinal layer segmentation.

Conclusions:

• The developed automated segmentation technique provides accurate retinal layer identification in OCT images.
• This method has the potential to improve computer-aided diagnosis for conditions like glaucoma.
• The technique offers a robust solution for analyzing ocular structures in OCT scans.