| Project/Area Number |
25350573
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Medical systems
|
|---|
| Research Institution | Kogakuin University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TOKUUE Kouichi 東京医科大学, 医学部放射線科, 主任教授 (00334061)
AKADA Souichi 東京医科大学, 医学部放射線科, 教授 (70246198)
|
|---|
| Co-Investigator(Renkei-kenkyūsha) |
MINORIKAWA Gaku 法政大学, 理工学部, 教授 (10318602)
|
|---|
| Research Collaborator |
KURONO Akihiko
KOIDE Tomomi
HIGUCHI SHiori
HISHIDA Keiko
MORISAKI Yutaka
YANAI Eisuke
|
|---|
| Project Period (FY) |
2013-04-01 – 2016-03-31
|
| Project Status |
Completed (Fiscal Year 2015)
|
| Budget Amount *help |
¥5,070,000 (Direct Cost: ¥3,900,000、Indirect Cost: ¥1,170,000)
Fiscal Year 2015: ¥130,000 (Direct Cost: ¥100,000、Indirect Cost: ¥30,000)
Fiscal Year 2014: ¥130,000 (Direct Cost: ¥100,000、Indirect Cost: ¥30,000)
Fiscal Year 2013: ¥4,810,000 (Direct Cost: ¥3,700,000、Indirect Cost: ¥1,110,000)
|
|---|
| Keywords | 人工知能 / ニューラルネットワーク / 心音 / 自己聴診 / 自動診断 / FFT / 固体差 / 雑音 / 個体差 / プログラム / 遠隔地診断 / 若手医師育成 |
|---|
| Outline of Final Research Achievements |
Trial study on an automated diagnosis system based on the processing of 2D-function on data envelope of the recorded heart sound has been completed. The expected apparatus is portable and usable at home. The system consists of recording components which catches and records a heart sound signal and stores it as WAVE data, a transforming component which converts the WAVE data to the TXT data to process the x-component for the frequency and the y-component for the time sequence with the z-component for the amplitude of heart sound data, and an artificial intelligence (AI) which gives a diagnosis based on the processing. Enlargement of a bell-size should be recommended for the better auscultation over clothes. Miniaturization of the apparatus is also necessarily taken into account. Data from 20 healthy persons showed large dependency on individuals, so that simplification of the information shown in the data envelope and more data would be of importance to learn the AI well.
|
