2014 Fiscal Year Final Research Report
Computer-Aided Diagnosis for Brain Check-up
| Project/Area Number |
24591815
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Radiation science
|
|---|
| Research Institution | Kumamoto University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2012-04-01 – 2015-03-31
|
| Keywords | ラクナ梗塞 / 未破裂脳動脈瘤 / コンピュータ支援診断 |
|---|
| Outline of Final Research Achievements |
The purpose of this study is to sophisticate two CAD schemes: (1) for the detection of lacunar infarcts on MR images and (2) for the detection of unruptured aneurysms on MRA images. To improve the method for distinguishing between lacunar infarcts and false positives, we developed eigenspace template matching, kernel eigenspace template matching, and AdaBoost template matching. In addition, we developed a novel technique by using gradient concentration filter bank and random forest in order to sophisticate the method for distinguishing between unruptured aneurysms and false positives. By using these proposed method, we could achieve a high detection performance compared with conventional methods.
|
| Free Research Field |
医用画像工学，コンピュータ支援診断
|
