| Project Area | Multidisciplinary computational anatomy and its application to highly intelligent diagnosis and therapy |
|---|
| Project/Area Number |
26108002
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area)
|
| Allocation Type | Single-year Grants |
|---|
| Review Section |
Science and Engineering
|
|---|
| Research Institution | Tokyo University of Agriculture and Technology |
|---|
Principal Investigator |
Shimizu Akinobu 東京農工大学, 工学(系)研究科(研究院), 教授 (80262880)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
縄野 繁 国際医療福祉大学, 医学部, 教授 (40156005)
小林 直樹 埼玉医科大学, 保健医療学部, 教授 (40523634)
庄野 逸 電気通信大学, 大学院情報理工学研究科, 教授 (50263231)
長谷川 巖 神奈川歯科大学, 歯学部, 教授 (00433912)
|
|---|
| Research Collaborator |
SAITO atsushi
LINGURARU George, Marius
SHINODA kazuma
KOMAGATA hideki
ISHIKAWA masahiro
|
|---|
| Project Period (FY) |
2014-07-10 – 2019-03-31
|
| Project Status |
Completed (Fiscal Year 2018)
|
| Budget Amount *help |
¥102,440,000 (Direct Cost: ¥78,800,000、Indirect Cost: ¥23,640,000)
Fiscal Year 2018: ¥19,240,000 (Direct Cost: ¥14,800,000、Indirect Cost: ¥4,440,000)
Fiscal Year 2017: ¥19,760,000 (Direct Cost: ¥15,200,000、Indirect Cost: ¥4,560,000)
Fiscal Year 2016: ¥19,370,000 (Direct Cost: ¥14,900,000、Indirect Cost: ¥4,470,000)
Fiscal Year 2015: ¥21,320,000 (Direct Cost: ¥16,400,000、Indirect Cost: ¥4,920,000)
Fiscal Year 2014: ¥22,750,000 (Direct Cost: ¥17,500,000、Indirect Cost: ¥5,250,000)
|
|---|
| Keywords | 計測工学 / 医用画像処理 / 解剖学 / 統計数理 / 画像 / 情報工学 / 統計数学 |
|---|
| Outline of Final Research Achievements |
Main research achievements are summarized as follows. First we proposed several methods to construct a spatio-temporal statistical model of time series data, such as statistical parameter estimation from a small sample data, modeling with smoothness constraint along a time axis, modeling with nested and neighboring constraints, and modeling of organs with topological changes along a time axis. These methods were used to construct spatio-temporal models of human embryos and children. Secondly, we proposed super-resolution techniques using dictionary learning and deep learning, and applied them to a CT volume. Third we reconstructed 3D tissue structure from pathological images using machine learning techniques. Fourth, hyper-spectral image processing of pathological images was conducted to improve accuracy in tissue classification. Fifth, we proposed a simultaneous optimization algorithm of segmentation and a statistical model, and used for organ recognition of a CT volume.
|
| Academic Significance and Societal Importance of the Research Achievements |
ヒト胚子の統計的変動の数理モデルは，臓器の発生へ対応しつつ，同時に，入れ子や隣接臓器間の重なりなどの構造上の制約を満たす必要がある．本研究では，これらの制約を満たす世界初の統計モデルを提案した．また，4次元の時空間において小児臓器の滑らかな変化を記述できる方法を提案し，小児のモデルの精度向上を達成した．超解像の研究では8倍の超解像に成功し，胸部CT像に適用して末梢の気管支が復元できることを示した．病理画像からの組織の3次元再構成では，機械学習とハイパースペクトル画像を用いて精度向上に取り組んだ．セグメンテーションと統計モデルの同時最適化では，世界で初めて現実的な時間で厳密解を導く方法を提案した
|
