• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Research on construction of highly accurate image recognition methods from limited supervised data

Research Project

Project/Area Number 19H01115
Research Category

Grant-in-Aid for Scientific Research (A)

Allocation TypeSingle-year Grants
Section一般
Review Section Medium-sized Section 61:Human informatics and related fields
Research InstitutionThe University of Tokyo

Principal Investigator

Harada Tatsuya  東京大学, 大学院情報理工学系研究科, 教授 (60345113)

Project Period (FY) 2019-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥44,460,000 (Direct Cost: ¥34,200,000、Indirect Cost: ¥10,260,000)
Fiscal Year 2022: ¥10,400,000 (Direct Cost: ¥8,000,000、Indirect Cost: ¥2,400,000)
Fiscal Year 2021: ¥14,950,000 (Direct Cost: ¥11,500,000、Indirect Cost: ¥3,450,000)
Fiscal Year 2020: ¥13,650,000 (Direct Cost: ¥10,500,000、Indirect Cost: ¥3,150,000)
Fiscal Year 2019: ¥5,460,000 (Direct Cost: ¥4,200,000、Indirect Cost: ¥1,260,000)
Keywords画像認識 / 機械学習 / 少数データ
Outline of Research at the Start

本研究の目的は,少数の教師情報しかない状況において,高精度な画像認識モデルを学習するアルゴリズムの構築である.そこで本提案では,1)限られた教師付きデータを最大限活用して深層学習の識別能力を可能な限り引き出す学習理論とアルゴリズム開発,2)シミュレーション等で学習したモデルをドメインの異なる実世界で動作させるためのドメイン適合手法の構築,3)効率的な教師データ作成のための深層学習に最適な能動学習アルゴリズムの開発,の3つの観点から,この困難な問題に取り組んでいく.

Outline of Final Research Achievements

Recent successes in deep learning have dramatically improved the accuracy of image recognition but achieving high recognition performance requires a huge amount of supervised data. Generating high-quality supervised data requires a lot of human effort and cost, which is a major problem in machine learning. In this study, we developed a method for learning highly accurate image recognition models with only a small amount of supervised data. Specifically, we developed a methodology to maximize the discriminative power of deep learning by making the most of limited supervised data, a domain adaptation method that enables knowledge transfer between different domains, and active information acquisition for efficient generation of supervised data.

Academic Significance and Societal Importance of the Research Achievements

現在成功している高精度の画像認識システムは教師あり学習を基盤としているが,大変なコストがかかるため機械学習分野において大問題となっている.さらに,付与するラベルに高度な専門知識を必要とする場合,アノテーションができる人が少数であり,膨大な教師データを作ることが不可能に近い.以上のように,教師データが入手困難な状況は多方面で存在する.従って,少数の教師データから高精度な画像認識モデルを学習するための方法論の実現は,現状の知的なシステムがより汎用的に利用されるための学術的,社会的最重要課題の一つであり,本研究成果はこの問題解決の一翼を担うものである.

Report

(6 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Annual Research Report
  • 2020 Annual Research Report
  • 2019 Comments on the Screening Results   Annual Research Report
  • Research Products

    (26 results)

All 2023 2022 2021 2020 2019 Other

All Journal Article (20 results) (of which Int'l Joint Research: 4 results,  Peer Reviewed: 19 results,  Open Access: 18 results) Presentation (3 results) (of which Int'l Joint Research: 3 results,  Invited: 3 results) Remarks (3 results)

  • [Journal Article] Backprop Induced Feature Weighting for Adversarial Domain Adaptation with Iterative Label Distribution Alignment2023

    • Author(s)
      Thomas Westfechtel, Hao-Wei Yeh, Qier Meng, Yusuke Mukuta, Tatsuya Harada
    • Journal Title

      Winter Conference on Applications of Computer Vision (WACV)

      Volume: - Pages: 392-401

    • DOI

      10.1109/wacv56688.2023.00047

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Non-rigid Point Cloud Registration with Neural Deformation Pyramid2022

    • Author(s)
      Yang Li, Tatsuya Harada
    • Journal Title

      Thirty-sixth Conference on Neural Information Processing Systems (NeurIPS)

      Volume: -

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Pop Music Generation with Controllable Phrase Lengths2022

    • Author(s)
      Daiki Naruse, Tomoyuki Takahata, Yusuke Mukuta, Tatsuya Harada
    • Journal Title

      The 23rd International Society for Music Information Retrieval Conference (ISMIR)

      Volume: -

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Unsupervised Pose-aware Part Decomposition for Man-made Articulated Objects2022

    • Author(s)
      Yuki Kawana, Yusuke Mukuta, Tatsuya Harada
    • Journal Title

      The European Conference on Computer Vision (ECCV)

      Volume: - Pages: 558-575

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Unsupervised Learning of Efficient Geometry-Aware Neural Articulated Representations2022

    • Author(s)
      Atsuhiro Noguchi, Xiao Sun, Stephen Lin, Tatsuya Harada
    • Journal Title

      The European Conference on Computer Vision (ECCV)

      Volume: - Pages: 597-614

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Deforming Radiance Fields with Cages2022

    • Author(s)
      Tianhan Xu, Tatsuya Harada
    • Journal Title

      The European Conference on Computer Vision (ECCV)

      Volume: - Pages: 159-175

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Lepard: Learning partial point cloud matching in rigid and deformable scenes2022

    • Author(s)
      Yang Li, Tatsuya Harada
    • Journal Title

      IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

      Volume: -

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Watch It Move: Unsupervised Discovery of 3D Joints for Re-Posing of Articulated Objects2022

    • Author(s)
      Atsuhiro Noguchi, Umar Iqbal, Jonathan Tremblay, Tatsuya Harada, Orazio Gallo
    • Journal Title

      IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

      Volume: -

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Boosting Source-free Domain Adaptation via Confidence-based Subsets Feature Alignment2022

    • Author(s)
      Hao-Wei Yeh, Thomas Westfechtel, Jia-Bin Huang, Tatsuya Harada
    • Journal Title

      26th International Conference on Pattern Recognition (ICPR)

      Volume: - Pages: 2857-2863

    • DOI

      10.1109/icpr56361.2022.9956719

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Neural Articulated Radiance Field2021

    • Author(s)
      Atsuhiro Noguchi, Xiao Sun, Stephen Lin, Tatsuya Harada
    • Journal Title

      International Conference on Computer Vision (ICCV)

      Volume: - Pages: 5762-5772

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Blur, Noise, and Compression Robust Generative Adversarial Networks2021

    • Author(s)
      Takuhiro Kaneko, Tatsuya Harada
    • Journal Title

      IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

      Volume: - Pages: 13579-13589

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Hyperbolic Neural Networks++2021

    • Author(s)
      Ryohei Shimizu, Yusuke Mukuta, Tatsuya Harada
    • Journal Title

      International Conference on Learning Representations (ICLR)

      Volume: -

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Real-Time Mesh Extraction from Implicit Functions via Direct Reconstruction of Decision Boundary2021

    • Author(s)
      Wataru Kawai, Yusuke Mukuta, Tatsuya Harada
    • Journal Title

      IEEE International Conference on Robotics and Automation (ICRA)

      Volume: -

    • Related Report
      2021 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Spherical Image Generation from a Single Image by Considering Scene Symmetry2021

    • Author(s)
      Takayuki Hara, Yusuke Mukuta, Tatsuya Harada
    • Journal Title

      Association for the Advancement of Artificial Intelligence (AAAI)

      Volume: -

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] RGBD-GAN: Unsupervised 3D Representation Learning From Natural Image Datasets via RGBD Image Synthesis2020

    • Author(s)
      Atsuhiro Noguchi, Tatsuya Harada
    • Journal Title

      2020 International Conference on Learning Representations

      Volume: -

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Bounding-box Channels for Visual Relationship Detection2020

    • Author(s)
      Sho Inayoshi, Keita Otani, Antonio Tejero-de-Pablos, Tatsuya Harada
    • Journal Title

      European Conference on Computer Vision

      Volume: - Pages: 682-697

    • DOI

      10.1007/978-3-030-58558-7_40

    • ISBN
      9783030585570, 9783030585587
    • Related Report
      2020 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Accurate Parts Visualization for Explaining CNN Reasoning via Semantic Segmentation2020

    • Author(s)
      Ren Harada, Antonio Tejero-de-Pablos and Tatsuya Harada
    • Journal Title

      The 31st British Machine Vision Conference

      Volume: -

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Noise Robust Generative Adversarial Networks2020

    • Author(s)
      Takuhiro Kaneko, Tatsuya Harada
    • Journal Title

      Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

      Volume: - Pages: 8404-8414

    • Related Report
      2020 Annual Research Report
    • Peer Reviewed / Open Access
  • [Journal Article] Label-Noise Robust Generative Adversarial Networks2019

    • Author(s)
      Takuhiro Kaneko, Yoshitaka Ushiku, Tatsuya Harada
    • Journal Title

      The 32nd IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR)

      Volume: - Pages: 2467-2476

    • DOI

      10.1109/cvpr.2019.00257

    • Related Report
      2019 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Class-Distinct and Class-Mutual Image Generation with GANs2019

    • Author(s)
      Takuhiro Kaneko, Yoshitaka Ushiku, Tatsuya Harada
    • Journal Title

      30th British Machine Vision Conference (BMVC)

      Volume: -

    • Related Report
      2019 Annual Research Report
  • [Presentation] Learning Deep Neural Networks from Limited Data2020

    • Author(s)
      Tatsuya Harada
    • Organizer
      GPU Technology Conference
    • Related Report
      2019 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Visual Recognition from Limited Supervised Data2019

    • Author(s)
      Tatsuya Harada
    • Organizer
      The 11th Asian Conference on Machine Learning (ACML)
    • Related Report
      2019 Annual Research Report
    • Int'l Joint Research / Invited
  • [Presentation] Domain Adaptation for Object Detection and Generation2019

    • Author(s)
      Tatsuya Harada
    • Organizer
      UK-Japan robotics and AI research collaboration workshops
    • Related Report
      2019 Annual Research Report
    • Int'l Joint Research / Invited
  • [Remarks] Tatsuya Harada

    • URL

      https://www.mi.t.u-tokyo.ac.jp/harada/

    • Related Report
      2022 Annual Research Report 2021 Annual Research Report 2020 Annual Research Report
  • [Remarks] rGAN

    • URL

      https://github.com/takuhirok/rGAN

    • Related Report
      2019 Annual Research Report
  • [Remarks] CP-GAN

    • URL

      https://github.com/takuhirok/CP-GAN

    • Related Report
      2019 Annual Research Report

URL: 

Published: 2019-04-18   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi