Automated Fundus Image Quality Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fundus image acquisition systems in remote or underserved areas often produce images of insufficient quality due to improper positioning, out-of-focus issues, or inadequate illumination, leading to delays in diagnosis and increased burden on patients, as quality assessment typically occurs after image capture rather than in real-time.
Innovation Solution
An automated method and system for fundus image field detection and quality assessment that identifies the eye and field of each image, evaluates image quality, and assigns a comprehensive quality metric, utilizing physiological landmarks like the optic disc and macula, and blood vessel symmetry, without requiring operator training or parameter setting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If automated quality assessment is implemented at image capture time, then image quality can be evaluated in real-time allowing immediate retaking, but the system complexity increases due to the need for automated field detection and quality metric calculation algorithms
Solution Approach 1:
The system performs field detection and quality assessment immediately after image capture, before the patient leaves the clinic. By executing the quality evaluation algorithm in real-time at the point of image acquisition, the system identifies poor quality images while the opportunity for immediate retaking still exists, thus eliminating the time delay inherent in traditional store-and-forward approaches where quality is assessed only after specialist review
Solution Approach 2:
The system uses automated algorithms to detect anatomical fields and assess image quality without requiring operator training or manual parameter setting. The computer automatically identifies the optic disc, macula, and other retinal structures, and calculates quality metrics based on image characteristics, making the system self-sufficient and reducing the need for trained personnel at the remote location
2Reliability
If multiple images are captured to ensure sufficient quality, then the likelihood of obtaining diagnostic images increases, but the time required for image acquisition and the burden on patients increases
Solution Approach 1:
The system performs comprehensive quality assessment of all captured images immediately after acquisition, before the patient leaves the clinic. By evaluating each image's quality metrics and determining whether the set meets diagnostic requirements while the patient is still present, the system identifies insufficient image sets early, allowing for targeted retaking of only necessary images rather than requiring multiple sequential visits
Solution Approach 2:
The system provides immediate feedback to the operator about the quality of captured images through automated quality metrics and sufficiency determination. This feedback loop allows the operator to understand which images meet quality standards and which require retaking, enabling informed decisions about whether additional images are needed and guiding the retaking process to efficiently achieve a sufficient image set
Data Source
AI summary
A method, system, and computer readable medium which automatically determine the side, field and a level of image quality of fundus images of the retina of a human eye is disclosed. The disclosure combines image processing, computer vision and pattern recognition techniques in a unique way to provide a robust process to identify and grade the quality of fundus images with application to improve efficiency and reduce errors in clinical or diagnostic retinal imaging workflows.