Development of an assisted diagnostic system for retinopathy that applies temporal subtraction to funduscopy images

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K01426
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Medical systems
• Research Institution: Kanagawa Institute of Technology
• Principal Investigator: TAKEO HIDEYA 神奈川工科大学, 工学部, 教授 (90434414)
• Project Period (FY): 2017-04-01 – 2020-03-31
• Keywords: 網膜症 / 眼底画像 / 経時差分画像 / コンピュータ支援画像診断 / 臨床評価

Outline of Final Research Achievements

We have developed a method of quantitatively expressing the degree of retinopathy symptoms by comparing two eye fundus images, one from the initial diagnosis and the other from follow-up. After conducting high-precision registration of the two images a differential image of the two registered images was created. Next, by calculating the sum of the square of the differential value from the differential image, a quantitative value was obtained that indicates symptomatic changes. In addition, focusing on the changes in the number of petechial hemorrhages that develop in the vicinity of the macula lutea, we weighted changes in numbers of petechial hemorrhages with the quantitative values derived from the differential values mentioned above.

Free Research Field

医用システム

Academic Significance and Societal Importance of the Research Achievements

本研究の独自性・創造性は，眼底画像を用いた糖尿病網膜症の主観的な診断において，経時差分画像と画像解析に基づく黄斑部周辺に生じた点状出血を考慮した定量値により診断に客観性を持たせていることであり，このようなアイデアによる研究は過去に例がなく独創的であると考える．
また，医師の主観や経験年数などの影響も軽減され，正診率が高まることは患者にとって非常に有益なことであることから，本研究の社会的意義は高いと考えられる．