| Project/Area Number |
15070203
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research on Priority Areas
|
| Allocation Type | Single-year Grants |
|---|
| Review Section |
Science and Engineering
|
|---|
| Research Institution | Gifu University |
|---|
Principal Investigator |
FUJITA Hiroshi Gifu University, Department of Intelligent Image Information, Division of Regeneration and Advanced Medical Sciences, Graduate School of Medicine, Professor (10124033)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
HOSHI Hiroaki Gifu University, School of Medicine, Department of Radiology, Professor (60128395)
HARA Takeshi Gifu University, Derpartment of Intelligent Image Information, Division of Regeneration and Advanced Medical Sciences, Graduate School of Medicine, Associate Professor (10283285)
ZHOU Xiangrong Gifu University, Department of Intelligent Image Information, Division of Regeneration and Advanced Medical Sciences, Graduate School of Medicine, Assistant Professor (00359738)
TSAI Du-yih Niigata University, School of Health Sciences, Professor (50178464)
KANEMATSU Masayuki Gifu University, School of Medicine, Department of Radiology, Associate Professor (40252134)
松原 友子 名古屋文理大学, 情報文化部, 助教授 (30279963)
福岡 大輔 岐阜工業高等専門学校, 助手 (60321436)
|
|---|
| Project Period (FY) |
2003 – 2006
|
| Project Status |
Completed (Fiscal Year 2007)
|
| Budget Amount *help |
¥56,000,000 (Direct Cost: ¥56,000,000)
Fiscal Year 2006: ¥14,000,000 (Direct Cost: ¥14,000,000)
Fiscal Year 2005: ¥14,000,000 (Direct Cost: ¥14,000,000)
Fiscal Year 2004: ¥14,000,000 (Direct Cost: ¥14,000,000)
Fiscal Year 2003: ¥14,000,000 (Direct Cost: ¥14,000,000)
|
|---|
| Keywords | Computer-aided diagnosis(CAD) / Image, character, sound recognition / Medical technology / 情報システム / 画像,文章,音声等認識 / 医用画像処理 / 医用画像認識 |
|---|
| Research Abstract |
We proposed a new type of computer-aided diagnosis (CAD) system to detect abnormalities based on the understanding of the normal anatomy structures of the human body, and concentrated on the developments on recognizing the anatomy structures of the human body automatically. The following study results were obtained. (1) A database including a large number of CT patient cases with a high image-resolution was constructed. Moreover, all information including the CT images, CT scan parameters, patient information, diagnosis reports, and the experiment results were built in electronically, and could be retrieved by the computer easily. (2) An automated scheme was developed and applied to 332 cases of torso CT images. The experimental results showed that the anatomical structures (skin, fat under skin, muscle, bone, visceral fat, lungs field, trachea, mediastinum, diaphragm and liver, etc. ) in about 70% cases were recognized correctly. Moreover, the doctor also confirmed the developed schem
… More
e could recognize the lung lobar structures successfully and generate the excellent results automatically. (3) An attempt to construct a human body model was carried out. A method was proposed for generation of the likelihood images for internal organs (liver, mammary gland, and muscle) based on the density and the position features. We confirmed the generated likelihood images were useful for the automated segmentations of internal organs and the judgment of lesions. In addition, the volumes of the woman mammary gland regions were measured in each age stage, and those results were used to construct a normal model of the mammary gland region. The volumes of the mammary gland regions measured by the proposed scheme could be expected to be used for prediction of the incidence of the breast cancer (risk evaluation) by the comparison with the normal model. (4) The results of the normal anatomical structure recognition were also used in a CAD system for automated detections of hemorrhage regions in head and abdominal contrast-enhanced CT images for emergency medical care, and the efficiency was confirmed. Less
|
