Digital Fundus Image Diagnostic System

Period of Performance: 09/30/1997 - 09/29/1998

Unknown

Phase 2 SBIR

Recipient Firm

Kestrel Corporation
3815 Osuna Road NE
Albuquerque, NM 87109
Principal Investigator

Abstract

A computer-based ophthalmic diagnostic system will be demonstrated and tested that will provide improved diagnosis and follow-up capabilities for the doctor as well as improved efficiencies for other clinic functions such as patient screening, examination preparation and record keeping. Personal computer-based image handling methodologies and image processing technologies will be embedded in a "Digital Fundus Image Diagnostic System" the first version of which will be commercialized near the end of the first year of Phase II. The system will provide a blend of unique image analysis techniques easily and inexpensively incorporated into existing fundus cameras with advanced image processing techniques. It will automate objective and quantitative analysis of fundus images taken over a period of time. The system will perform an objective, computer-based analyses of features associated with diabetic retinopathy. The research will compare computer-based analysis techniques with results from human technicians. The analysis to measure how lesions are changing in shape or size, detect neovascularization, etc. will be fully automated. PROPOSED COMMERCIAL APPLICATION: Approximately 15,000 ophthalmologists, practicing in eye care clinics that form the core of eye care providers in the U.S, are becoming increasingly consolidated into large healthcare partnerships and organizations to focus on more efficient and cost- effective means of providing quality patient care. Imaging technology- based integrated systems that combine voice recognition systems and multimedia data storage and transfer capabilities, with pattern recognition-based, computer-assisted diagnosis and image enhancement capabilities will lead to more timely and better diagnoses and more efficient storage, retrieval and availability of patient records.