Industrial Technology Research Institute

Technology Overview

The AI Decision Support System of Fundus Image is a collaboration of the minds of the human brain and artificial intelligence through the expertise of ophthalmologists and AI analysis. Composed with complementary AI models, it can perform individual diagnostic tasks like doctors, such as categorizing and detecting, improve the efficacy of overall interpretation. Therefore, this innovation increases the early detection rate of diabetic retinopathy (DR) for potential patients and reduces the cost of healthcare and social care. In addition, the combination of hardware and software into a portable medical device makes it easy to carry around and can shorten the time to diagnosis. This portable solution with AI edge computing technology is jointly developed by Industrial Technology Research Institute (ITRI) and Medimaging Integrated Solution Inc. (MiiS).

• Leading AI-assisted Diagnosis System
  • Currently the only AI detection technology in the world that can detect the four main lesions (Microaneurysms, Hemorrhages, Soft Exudates, Hard Exudates) in DR, label the corresponding locations clearly, and assist doctors in interpreting the severity of the condition. (Figure 2)
  • In addition to diagnosis, with the ability to filter poor-quality images caused by dirt, shadows, and eyelashes, this technology can remind doctors or technicians to retake the image and accordingly prevent misinterpretation by artificial intelligence.
• Portable Solution & SaMD Clearance
  • Collaborated with the medical device designer and manufacturer 'Medimaging Integrated Solution Inc. (MiiS)' to develop the portable solution, including the handheld fundus camera and AI edge computer.
  • The world’s first AI-DR handheld fundus camera, which has been cleared by the TFDA as a Class II medical device providing both CADx and CADe, was launched in Taiwan in July 2021. (Figure 1)

Through intensification of the AI solution, this technology has extended the capability to simultaneously detect both diabetic retinopathy and diabetic macular edema (DME) and currently focuses on other 10+ ocular fundus abnormalities common to the general public.

Applications & Benefits

Portable Edge AI-DR uses AI to assist non-ophthalmologists in their diagnoses and solve the problem of uneven distribution and lack of ophthalmologists. By identifying DR stages and lesions, this innovation improves diabetes health monitoring and management and helps lower the risks of vision loss or blindness caused by DR.

• Innovative Diabetes Care Scenario
  • By helping non-ophthalmologists comprehensively conduct the detection of subtle lesions in fundus images, this innovation eliminates the inconvenience of referrals to ophthalmologists for patients without more than mild DR (mtmDR). (Figure 3)
  • Physicians can effectively grasp the patient's diabetes development status by tracking the changes in fundus symptoms over time. This is crucial to diabetes management as the changes in the fundus symptoms reflect better than blood sugar levels tested, which patients can control right before hospital revisits.
  • Beneficial for non-ophthalmologists and optometrists to provide advanced care for diabetic eye complications, this competitive AI-aided system and evolutionary service scenario will create new market opportunities for the global healthcare industry.
• Social Responsibility and Impact
  • Applicable for both community and hospital settings, the mobile service provided by Portable Edge AI-DR effectively assists in DR screening and early intervention and thus greatly reduces medical costs.
  • In collaboration with local clinics, Lions clubs, and medical centers, the system has been used for eye examinations for over 60,000 diabetics diabetics in Taiwan and successfully promoted a significant increase in DR screening rate in counties, e.g. a 27% increase in Taitung. (Figures 4 & 5)