Automatic Detection of Good/Bad Colonies of iPS Cells Using Deep Learning

• Project/Area Number: 16K00394
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Life / Health / Medical informatics
• Research Institution: Hiroshima University
• Principal Investigator: Raytchev Bisser 広島大学, 工学研究科, 准教授 (00531922)
• Project Period (FY): 2016-04-01 – 2019-03-31
• Project Status: Completed (Fiscal Year 2018)
• Budget Amount: ¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000); Fiscal Year 2018: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000); Fiscal Year 2017: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2016: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
• Keywords: iPS 細胞 / 機械学習 / ニューラルネットワーク / 深層学習 / セグメンテーション / 異常検出 / iPS cells / Machine learning / Neural networks / deep learning / segmentation / CNN / colony detection / iPS cell colonies / 画像認識

Outline of Final Research Achievements

In order to fulfill their promise in regenerative medicine and drug discovery, a steady supply of iPS cells obtained through harvesting of individual cell colonies is needed. However, cultivating iPS cell colonies is a sensitive process, and even if care is taken abnormalities can appear, which need to be detected. It is therefore important to automate the process of detecting such abnormalities and in this research project we have designed a new deep learning based semantic segmentation algorithm which is able to automatically detect and label at pixel level differentiated vs. undifferentiated cells in iPS colonies images.

Academic Significance and Societal Importance of the Research Achievements

再生医療の実現に向けて需要が高まっているiPS細胞の大量生産を可能にするため，培養を自動化する過程で必要になるiPS細胞の分化・未分化の自動検出ができる新しいアルゴリズムを開発して，その有用性を検証した．また，本研究で開発したSemantic segmentation アルゴリズムが従来手法より高精度であるため，他の医用画像解析への応用も期待できる．

Research Products

• [Journal Article] Detection of Differentiated vs. Undifferentiated Colonies of iPS Cells Using Random Forests Mod-eled with the Multivariate Polya Distribution (2016)
  • Author(s): B. Raytchev, A. Masuda, M. Minakawa, K. Tanaka, T. Kurita, T. Imamura, M. Suzuki, T. Tamaki and K. Kaneda
  • Journal Title: Springer Lecture Notes in Computer Science (LNCS) Volume: 9901 Pages: 667-675
  • DOI: 10.1007/978-3-319-46723-8_77
  • ISBN: 9783319467221, 9783319467238
  • Peer Reviewed / Acknowledgement Compliant
• [Presentation] Patch-based learning for biomedical image analysis (2018)
  • Author(s): Bisser Raytchev
  • Organizer: Hiroshima Conference on Statistical Science 2016
  • Int'l Joint Research / Invited
• [Presentation] Detection of Differentiated vs. Undifferentiated Colonies of iPS Cells Using Random Forests Modeled with the Multivariate Polya Distribution (2018)
  • Author(s): Bisser Raytchev
  • Organizer: 第20回 画像の認識・理解シンポジウム(MIRU2017)
  • Invited
• [Presentation] 構造情報を用いた分類型CNNによるiPS細胞の分化・未分化検出 (2018)
  • Author(s): 林 祥平, Bisser Raytchev, 玉木 徹, 金田 和文
  • Organizer: 第21回 画像の認識・理解シンポジウム(MIRU2018)
• [Presentation] Multi-Scale Scene Recognition Using Small Training Datasets (2018)
  • Author(s): Tokinirina Radiniaina, Bisser Raytchev, Kazufumi Kaneda, Toru Tamaki
  • Organizer: 第21回 画像の認識・理解シンポジウム(MIRU2018)
• [Presentation] 回帰型CNNを用いたiPS細胞の分化・未分化検出 (2018)
  • Author(s): 田中孝二郎, ライチェフ ビせル, 栗田多喜夫, 玉木徹, 金田和文
  • Organizer: 電子情報通信学会 パターン認識・メディア理解研究会（PRMU）技術報告 PRMU2017-164, vol. 117, no. 442, pp. 109-114
• [Presentation] Grassmann Matching Kernels for Scene Representation and Recognition (2017)
  • Author(s): Bisser Raytchev, Miku Koujiba, Toru Tamaki and Kazufumi Kaneda
  • Organizer: International Joint Conference on Neural Networks (IJCNN)
  • Place of Presentation: William A. Egan Civic and Convention Center in Anchorage, Alaska, USA
  • Year and Date: 2017-05-14
  • Int'l Joint Research
• [Presentation] Grassmann Matching Kernels for Scene Representation and Recognition (2017)
  • Author(s): Bisser Raytchev, Miku Koujiba, Toru Tamaki and Kazufumi Kaneda
  • Organizer: International Joint Conference on Neural Networks (IJCNN 2017), pp. 4422-4428, Anchorage, USA
  • Int'l Joint Research
• [Presentation] Ensemble-Based Local Learning for High-Dimensional Data Regression (2016)
  • Author(s): B. Raytchev, Y. Katamoto, M. Koujiba, T. Tamaki and K. Kaneda
  • Organizer: 23rd International Conference on Pattern Recognition (ICPR)
  • Place of Presentation: Cancun Center, Cancun, Mexico
  • Year and Date: 2016-12-04
  • Int'l Joint Research
• [Presentation] Higher-Level Representation of Local Spatio-Temporal Features for Human Action Recognition Using Subspace Matching Kernels (2016)
  • Author(s): B. Raytchev, H. Kawamoto, T. Tamaki and K. Kaneda
  • Organizer: 23rd International Conference on Pattern Recognition (ICPR)
  • Place of Presentation: Cancun Center, Cancun, Mexico
  • Year and Date: 2016-12-04
  • Int'l Joint Research
• [Presentation] 深層学習を用いた大腸NBI内視鏡画像認識 (2016)
  • Author(s): 田中 孝二郎, Bisser Raytchev, 玉木 徹, 小出 哲士, 吉田 成人, 三重野 寛, 田中 信治
  • Organizer: 第19回画像の認識・理解シンポジウム(MIRU2016)
  • Place of Presentation: アクトシティ浜松, 静岡
  • Year and Date: 2016-08-01
• [Presentation] Detection of iPS Cell Colonies with Local Features (2016)
  • Author(s): Atsuki Masuda, Bisser Raytchev, Takio Kurita,Toru Imamura, Masashi Suzuki, Toru Tamaki and Kazufumi Kaneda
  • Organizer: 第35回日本医用画像工学会大会(JAMIT 2016)
  • Place of Presentation: 千葉大学西千葉キャンパス, 千葉
  • Year and Date: 2016-07-21