Application of deep learning and GMFH-type neural network for medical image diagnosis and affective engineering
| Project/Area Number |
18K04206
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 21040:Control and system engineering-related
|
|---|
| Research Institution | The University of Tokushima |
|---|
Principal Investigator |
TAKAO Shoichiro 徳島大学, 大学院医歯薬学研究部(医学域), 准教授 (30363146)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
上野 淳二 徳島大学, 大学院医歯薬学研究部(医学域), 非常勤講師 (60116788)
近藤 正 徳島大学, 大学院医歯薬学研究部(医学域), 名誉教授 (80205559)
|
|---|
| Project Period (FY) |
2018-04-01 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2020: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Fiscal Year 2019: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2018: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
|
|---|
| Keywords | 深層学習 / 医用画像認識 / GMDH型人工知能技術 / 医用画像診断 / 感性工学 |
|---|
| Outline of Final Research Achievements |
In this study, the deep hybrid neural networks which are constructed with the deep Group Method of Data Handling (GMDH)-type neural network and the convolutional neural network (CNN), is developed and these hybrid algorisms are applied to the medical image analysis of MRI images of the brain regions, the medical image analysis of X-ray CT images of the chest and the abdominal regions. The deep GMDH-type neural network can automatically organize the optimal deep neural network architectures. In this study, the deep neural networks which are used to recognize many organs, are automatically organized using the deep hybrid neural networks, and the recognition results are compared with the results of the conventional three-layered neural networks and it is shown that the hybrid neural networks are useful for the medical image recognitions of many organs.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究では、ディープGMDH-typeニューラルネットワークとCNNを組み合わせたハイブリッド型ニューラルネットワークを開発した。ディープGMDH-typeニューラルネットワークは、多くの中間層を持つニューラルネットワーク構造を自動的に自己組織でき、いろいろな臓器の医用画像認識問題に対して、最適な複雑さをしたネットワーク構造を自動的に自己組織できる。このような機能を備えた機械学習の数学アルゴリズムは他にない。本研究では、頭部、胸部、腹部のいろいろな臓器に対してハイブリッド型アルゴリズムを適用してその有効性を確認した。本アルゴリズムは汎用性が高く、他の分野にも簡単に応用が可能である。
|
Report
(4 results)
Research Products
(10 results)
