多重線形スパースと低ランクテンソル分解法による４次元生物データの解析

• Project Area: Resonance Biology for Innovative Bioimaging
• Project/Area Number: 18H04747
• Research Category: Grant-in-Aid for Scientific Research on Innovative Areas (Research in a proposed research area)
• Allocation Type: Single-year Grants
• Research Institution: Ritsumeikan University
• Principal Investigator: 陳 延偉 立命館大学, 情報理工学部, 教授 (60236841)
• Project Period (FY): 2018-04-01 – 2020-03-31
• Project Status: Completed (Fiscal Year 2019)
• Budget Amount: ¥5,200,000 (Direct Cost: ¥4,000,000、Indirect Cost: ¥1,200,000); Fiscal Year 2019: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000); Fiscal Year 2018: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
• Keywords: 4次元顕微鏡画像 / 多重線型理論 / スペースコーディング法 / 深層学習 / 細胞分裂検出 / ダイナミックCT画像 / 肝臓腫瘍分類 / 4Dバイオイメージ / 機械学習 / 多重線形解析 / 低ランクテンソル / 4次元画像 / スパース符号化法 / 深層学習法 / 2.5D 畳込ニューラルネットワーク / ４次元生物データ / スパース / 低ランク / 細胞分裂の検出

Outline of Annual Research Achievements

イメージング技術の飛躍的な進展により, 高精細なバイオメディカル画像情報が得られるようになってきた. バイオメディカル画像は３次元空間情報に加え, 動的時間情報ももつ．一種の４次元画像である. また, ラベリングには専門家の知識が必要である．一般画像解析に比べ，バイオメディカル画像解析では高次元での処理や学習用サンプルが少ないなどのチャレンジ課題が多い．本研究では，従来の機会学習法を多重線型代数の枠組で拡張し，４次元バイオメディカル画像を一つのテンソルとして時空間特徴を同時に抽出する手法を開発し、４次元バイオメディカルイメージの解析を効率よくできるようになった. さらに，近年注目されている深層学習法を用いた４次元バイオメディカル画像解析法も開発し, 従来の機会学習法より高精度な解析ができるようになった．主な成果は以下に示す．
(1) 多重線形スパースコーディング法を開発し，ダイナミックCT画像（４次元CT画像）から時空間特徴の抽出法を開発し、肝臓腫瘍鑑別法を開発した.
(2) 2.5D CNNとRNNを組み合わせたモデルを開発し，４次元顕微鏡画像から時空間特徴抽出法を開発し，高精度で細胞分裂を検出することができた.
これらの成果は, 国際学術誌Pattern Recognition Letter (impact factor: 2.81), Neural Computing and Application (Impact factor:4.664), IEEE/ACM Transactions on Computational Biology and Bioinformatics (Impact factor: 2.4), 及びトップ国際学会IEEE ICASSP2019 (H5 index: 70)などで発表してきた.

Research Progress Status

令和元年度が最終年度であるため、記入しない。

Strategy for Future Research Activity

令和元年度が最終年度であるため、記入しない。

Research Products

• [Journal Article] An end-to-end CNN and LSTM network with 3D anchors for mitotic cell detection in 4D microscopic images and its parallel implementation on multiple GPUs (2020)
  • Author(s): T. Kitrungrotsakul, X.-X Han, Y. Iwamoto, S. Takemoto, H. Yokota, S. Ipponjima, T. Nemoto, X. Wei, and Y.-W. Chen,
  • Journal Title: Neural Computing and Applications Volume: Vol. 32 Issue: 10 Pages: 5669-5679
  • DOI: 10.1007/s00521-019-04374-8
  • Peer Reviewed / Open Access
• [Journal Article] A Cascade of 2.5D CNN and Bidirectional CLSTM Network for Mitotic Cell Detection in 4D Microscopy Image (2019)
  • Author(s): Titinunt Kitrungrotsakul, Yutaro Iwamoto, Xian-Hau Han, Satoko Takemoto, Hideo Yokota, Sari Ipponjima, Tomomi Nemoto, Xiong Wei, Yen-Wei Chen
  • Journal Title: proc. IEEE/ACM Transactions on Computational Biology and Bioinformatics Volume: － Issue: 2 Pages: 1-9
  • DOI: 10.1109/tcbb.2019.2919015
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Journal Article] Tensor-based Sparse Representations of Multi-phase Medical Images for Classification of Focal Liver Lesions (2019)
  • Author(s): Jian Wang, Jing Li, Xian-Hua Han, Lanfen Lin, Hongjie Hu, Yingying Xu, Qingqing Chen, Yutaro Iwamoto, Yen-Wei Chen
  • Journal Title: Pattern Recognition Letter Volume: 印刷中 Pages: 207-215
  • DOI: 10.1016/j.patrec.2019.01.001
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Journal Article] Three-Dimensional Embryonic Image Segmentation and Registration Based on Shape Index and Ellipsoid-fitting Method (2019)
  • Author(s): Sihai Yang, Xianhua Han and Yen-Wei Chen
  • Journal Title: International Journal of Computational Biology Volume: 26 Issue: 2 Pages: 128-142
  • DOI: 10.1089/cmb.2018.0165
  • Peer Reviewed / Open Access
• [Presentation] Tensor-based Multilinear Subspace Learning for Multi-Dimensional Biomedical Image Analysis (2019)
  • Author(s): Yen-Wei Chen
  • Organizer: 2019 2nd International Conference on Image and Video Processing (ICIVP2019)
  • Int'l Joint Research / Invited
• [Presentation] Tensor Sparse Coding for Multi-Dimensional Medical Image Analysis (2019)
  • Author(s): Yen-Wei Chen
  • Organizer: 2019 2nd International Conference on Digital Medicine and Image Processing (DMIP2019)
  • Int'l Joint Research / Invited
• [Presentation] Computational Anatomy, Artificial Intelligence and Virtual Reality in Medicine and Healthcare (2019)
  • Author(s): Yen-Wei Chen
  • Organizer: The 7th IEEE International Conference on Serious Games and Applications for Health (IEEE SeGAH 2019)
  • Int'l Joint Research / Invited
• [Presentation] A Cascade of CNN and LSTM Network with 3D Anchors for Mitotic Cell Detection in 4D Microscopic Image (2019)
  • Author(s): Titinunt Kitrungrotsakul, Yutaro Iwamoto, Xian-Hua Han, Satoko Takemoto, Hideo Yokota, Sari Ipponjima, Tomomi Nemoto, Xiong Wei, Yen-Wei Chen
  • Organizer: IEEE ICASSP 2019
  • Int'l Joint Research
• [Presentation] Tensor-Based Subspace Learning for Classification of Focal Liver Lesions in Multi-phase CT Images (2019)
  • Author(s): Jian Song, Sihang Zhu, Lanfen Lin, Hongjie Hu, Yen-Wei Chen
  • Organizer: ICNC-FSDK 2019
  • Int'l Joint Research
• [Presentation] Artificial Intelligence and Virtual Reality in Medical Applications (2018)
  • Author(s): Yen-Wei Chen
  • Organizer: 2018 3rd International Conference on Information and Network Technologies (ICINT 2018)
  • Int'l Joint Research / Invited
• [Presentation] A 2.5D Cascaded Convolutional Neural Network with Temporal Information for Automatic Mitotic Cell Detection in 4D Microscopic Images (2018)
  • Author(s): Titinunt Kitrungrotsakul, Xian-Hau Han, Yutaro Iwamoto, Satoko Takemoto, Hideo Yokota, Sari Ipponjima, Tomomi Nemoto, Xiong Wei and Yen-Wei Chen
  • Organizer: The 14th International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC- FSKD2018)
  • Int'l Joint Research
• [Presentation] Focal Liver Lesion Classification Based on Tensor Sparse Representations of Multi-Phase CT Images (2018)
  • Author(s): Jian Wang, Xian-Hua Han, Jiande Sun, Lanfen Lin, Hongjie Hu, Yingying Xu, Qingqing Chen, and Yen-Wei Chen
  • Organizer: The 2018 Pacific-Rim Conference on Multimedia (PCM2018)
  • Int'l Joint Research