Corneal Epithelium Segmentation in OCT Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
OCT images suffer from speckle noise and other interferences, making it challenging to accurately segment the corneal epithelium layer due to low contrast and thin boundary layers, and the lack of sufficient annotated image data hinders the training of effective image segmentation models.
Innovation Solution
A method is described that involves generating a binarized image and a binary mask of the cornea from an OCT image, segmenting the anterior cornea and Bowman's layer, and using these segments to generate an epithelium map, all without requiring extensive training on large datasets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image segmentation methods are used on OCT images, then the segmentation process can be performed, but the accuracy is reduced due to speckle noise and low contrast in the corneal epithelium layer
Solution Approach 1:
The patent applies preliminary image processing steps including speckle noise reduction filtering and contrast enhancement before segmentation. These preprocessing actions prepare the OCT image by reducing harmful speckle noise and improving the visibility of the corneal epithelium layer boundaries, thereby enabling more accurate subsequent segmentation without requiring extensive training data
2Measurement precision
If machine learning models are trained with sufficient annotated image data to achieve high accuracy, then segmentation performance improves, but the requirement for large amounts of training data and compute resources increases
Solution Approach 1:
The patent performs preliminary image processing (speckle noise reduction and contrast enhancement) to improve image quality before segmentation, which enables accurate segmentation without requiring large annotated datasets for training machine learning models
Solution Approach 2:
The patent introduces intermediate processing steps including speckle noise reduction filtering and contrast enhancement as mediators between the raw OCT image and the final segmentation. These intermediate processing stages improve the quality of input data for segmentation algorithms, reducing the need for extensive training data while maintaining high accuracy
3Measurement precision
If extensive training data and compute resources are used to train segmentation models, then model accuracy improves, but the training time and computational cost increase
Solution Approach 1:
The patent applies speckle noise reduction and contrast enhancement preprocessing to OCT images before segmentation, which improves image quality and enables accurate segmentation without requiring time-consuming training of machine learning models on large datasets
Solution Approach 2:
The patent replaces the mechanical approach of training machine learning models with extensive data and compute resources with an alternative approach using signal processing techniques (speckle noise reduction filtering and contrast enhancement) to achieve accurate segmentation more efficiently
Data Source
AI summary
The techniques described herein provide improved techniques for segmenting corneal epithelium layer. A method includes receiving an optical coherence tomography (OCT) image of an eye; generating, based on the OCT image, a binarized image of the eye; generating, based on the binarized image of the eye and the OCT image, a binary mask of a cornea of the eye; segmenting, based on the binary mask of the cornea of the eye, an anterior cornea of the eye on the OCT image; generating, based on the OCT image and the segmented anterior cornea, a binary mask for an epithelium layer of the eye; segmenting, based on the binary mask for the epithelium layer of the eye, a Bowman's layer in the cornea of the eye on the OCT image; and causing the segmented anterior cornea and the segmented Bowman's layer data to be used for generation of an epithelium map.