Sponsored By

Technology Offers New Way of Seeing Blindness

Originally Published MDDI May 2005R&D DIGEST

Maria Fontanazza

May 1, 2005

2 Min Read
Technology Offers New Way of Seeing Blindness

Originally Published MDDI May 2005


Maria Fontanazza

Visually similar images captured by content-based image technology (top and middle) may enable researchers, including Edward Chaum (botom), to create a computer diagnosis system for early-stage eye diseases.

Researchers at the Oak Ridge National Laboratory (ORNL; Oak Ridge, TN) are working to develop a database to help diagnose eye ailments leading to blindness. The database combines image retrieval technology with a method of image analysis.

The merging technologies could also have uses in biomedical imaging and telemedicine.

The new diagnostic method could prevent or treat nearly 80% of blindness cases, says Edward Chaum, MD, PhD, one of the research team members. About 180 million people are blind or at risk for blindness.

“The innovation is not in the retinal imaging system per se,” says Chaum, an associate professor of ophthalmology and other fields at the University of Tennessee Health Science Center (Memphis). “Rather, it's the computer-based image analysis approach and the development of the image retrieval algorithms.” The technique is a new application of a technology called content-based image retrieval. It's been used to identify and categorize computer chip defects for the semiconductor industry. Visually similar images are located and sorted in a database. Merging analysis with image collection tools produces a computer diagnosis system. Applied to retinal shapes, it could detect potentially blinding eye diseases such as glaucoma or macular degeneration.

The team hopes to use optical coherence tomography and digital retinal photography to determine disease-related retinal changes. The database would contain images of recognized retinal diseases. Further testing is needed before the system is finalized, however.

“We are currently working on optimizing the diagnostic algorithms for specific retinal and optic nerve diseases and testing these in clinical studies to determine the sensitivity and specificity of the approach,” says Chaum. “Our success—or lack thereof—with the clinical trials will then direct us toward improvements in the algorithms and expansion of the clinical database.”

ORNL's Laboratory Directed Research and Development Program is funding the project.

Copyright ©2005 Medical Device & Diagnostic Industry

Sign up for the QMED & MD+DI Daily newsletter.

You May Also Like