Automatic Detection and Diagnostics of Diabetic Retinopathy
Background: Diabetes is a condition that affects blood vessels throughout the body, particularly in the kidneys and eyes. Diabetic retinopathy is a complication of diabetes and is the leading cause of new blindness in the United States. Diabetic retinopathy results when diabetes affects vessels in the eyes, producing abnormalities such as microaneurysms and hemorrhages. These abnormalities are the same color as that of blood vessels, causing some areas of the normal blood vessel system in the retina to be erroneously classified as defects. Automated systems for detecting diabetic retinopathy have been plagued by high false alarm rates. Performance reported from prior systems using a matched filter response was below the level of a human operator. Technology: UC researchers have developed a tool to aid in the automated detection of diabetic retinopathy. This novel system provides for blood vessel detection in a clutter (abnormality) environment and allows for efficient recognition of low-contrast minor blood vessels. A computer algorithm using adaptive thresholding and the localized Radon transform detects the normal vascular network from a retinal image. The detection system possesses overall superior performance than prior reported systems, and is simpler and less computationally expensive. Unlike prior systems, this method performs almost identically for both normal and abnormal cases and reduces the false alarm rate in detection of abnormalities. Application: The present invention is particularly useful for effective recognition of retinal vasculature among abnormalities. Additionally, the system can be used to detect blood vessels in general as part of a comprehensive diagnostic biomedical tool.
Type of Offer:
« More Biotech Patents« More Software Patents