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

Object detection using distortion-tolerant, controllable, parametric image matching

Research Project

Project/Area Number 17K00250
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Research Field Perceptual information processing
Research InstitutionHosei University

Principal Investigator

WAKAHARA Toru  法政大学, 情報科学部, 教授 (40339510)

Project Period (FY) 2017-04-01 – 2020-03-31
Project Status Completed (Fiscal Year 2019)
Budget Amount *help
¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
Fiscal Year 2019: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2018: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2017: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000)
Keywordsパターン認識 / 変形耐性画像マッチング / 物体検出 / 変形耐性 / 画像マッチング / コンピュータビジョン
Outline of Final Research Achievements

1. A combination of projection transformation tolerant GPT correlation method and multi-scale subwindow search method achieved a high accuracy of area-based object detection. 2. Enhanced GAT and GPT correlation methods using weights based on similarity in HOG features between corresponding points improved both convergence speed and matching ability. 3. Instead of using time-consuming multi-scale subwindow search method, we proposed a high-speed technique of search for optimal, initial matching location by means of a combination of sliding discrete Fourier transformation and Fourier descriptor of histogram of edge directions. 4. Reformulation of GAT and GPT correlation methods so that they should guarantee the L2 norm of template image in a rigorous manner improved greatly an ability of image matching.

Academic Significance and Societal Importance of the Research Achievements

本研究は、物体検出のための画像マッチングの研究分野で、現在主流でありながら十分な性能が達成できていない特徴ベースのアプローチではなく、変形耐性画像マッチングを用いた領域ベースの物体検出を行うという独特かつ明快な研究戦略に立つ。本研究の最たる独創性は、物体検出で原理的に要請される変形耐性をあらかじめパラメトリックな変形モデルとして搭載したGAT相関法およびGPT相関法をベースとし,テンプレート画像の全体と目標画像中の部分領域との間で最適な「全体-部分」マッチングを追求する点にある。計算量の大幅削減による高速化を得てその実用可能性が明らかになれば、本研究の社会的意義ならびに波及効果は極めて大きい。

Report

(4 results)
  • 2019 Annual Research Report   Final Research Report ( PDF )
  • 2018 Research-status Report
  • 2017 Research-status Report
  • Research Products

    (7 results)

All 2020 2019 2018 2017

All Journal Article (6 results) (of which Int'l Joint Research: 1 results,  Peer Reviewed: 6 results,  Open Access: 2 results) Presentation (1 results)

  • [Journal Article] Initial Location Search Method using Target Region Features for Region-Based Image Matching2020

    • Author(s)
      山下幸彦、若原徹
    • Journal Title

      電子情報通信学会論文誌D 情報・システム

      Volume: J103-D Issue: 4 Pages: 330-340

    • DOI

      10.14923/transinfj.2019JDP7056

    • ISSN
      1880-4535, 1881-0225
    • Year and Date
      2020-04-01
    • Related Report
      2019 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Image Alignment Using Norm Conserved GAT Correlation2019

    • Author(s)
      Toru Wakahara, Yukihiko Yamashita
    • Journal Title

      Proc. of 2019 International Conference on Digital Image Computing: Techniques and Applications

      Volume: 1 Pages: 48-52

    • DOI

      10.1109/dicta47822.2019.8945880

    • Related Report
      2019 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Practice in Caption Generation with Keras: The Design and Evaluation for Attention Models2019

    • Author(s)
      Rong Wang, Toru Wakahara
    • Journal Title

      Proc. of 3rd International Conference on Deep Learning Technologies

      Volume: 1 Pages: 11-15

    • DOI

      10.1145/3342999.3343004

    • Related Report
      2019 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Image registration using 2D projection transformation invariant GPT correlation2019

    • Author(s)
      Toru Wakahara, Shizhi Zhang, Yukihiko Yamashita
    • Journal Title

      Proc. of 2019 International Workshop on Advanced Image Technology (IWAIT2019)

      Volume: 11049 Pages: 5-5

    • DOI

      10.1117/12.2517185

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Binarization of color character strings in scene images using deep neural network2018

    • Author(s)
      Wengiao Bian, Toru Wakahara, Tao Wu, He Tang, Jirui Lin
    • Journal Title

      Proc. of Digital Image Computing: Techniques and Applications (DICTA2018)

      Volume: 1 Pages: 1-6

    • DOI

      10.1109/dicta.2018.8615837

    • Related Report
      2018 Research-status Report
    • Peer Reviewed / Open Access / Int'l Joint Research
  • [Journal Article] Image Matching Using GPT Correlation Associated with Simplified HOG Patterns2017

    • Author(s)
      Shizhi Zhang, Toru Wakahara, Yukihiko Yamashita
    • Journal Title

      Proc. of 7th International Conference on Image Processing Theory, Tools and Applications

      Volume: 1 Pages: 1-6

    • DOI

      10.1109/ipta.2017.8310122

    • Related Report
      2017 Research-status Report
    • Peer Reviewed
  • [Presentation] 領域画像マッチングのための初期探索法2018

    • Author(s)
      山下幸彦、若原徹
    • Organizer
      電子情報通信学会PRMU研究会
    • Related Report
      2018 Research-status Report

URL: 

Published: 2017-04-28   Modified: 2021-02-19  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi