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

Developing Semantic Image Synthesis Model Using Limited Training Data

Research Project

Project/Area Number 20K19816
Research Category

Grant-in-Aid for Early-Career Scientists

Allocation TypeMulti-year Fund
Review Section Basic Section 61010:Perceptual information processing-related
Research InstitutionUniversity of Tsukuba

Principal Investigator

Endo Yuki  筑波大学, システム情報系, 助教 (00790396)

Project Period (FY) 2020-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000)
Fiscal Year 2022: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2021: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
Fiscal Year 2020: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
Keywords深層学習 / 画像生成 / GAN / 意味的画像合成 / 畳み込みニューラルネットワーク / GAN inversion / コンピュータグラフィックス / コンピュータビジョン / 変分オートエンコーダ / 画像合成
Outline of Research at the Start

意味的画像合成は、ユーザがキャンバス上に「ここは建物、ここは木」という風に粗くラベルを塗るだけで、直感的に画像を生成できる技術であり、世界的に注目を集めている。深層学習によって大量の教師データを用いた学習をすれば写実的な画像を作れるが、教師データ作成の人的コストは大きい。本研究では、限られたラベル付教師データやラベルなし訓練データを効率的に活用できる深層学習の枠組みを開拓し、従来よりも高品質な意味的画像合成の実現を目指す。

Outline of Final Research Achievements

Semantic image synthesis is a technique that can generate images from a semantic map annotated with pixel-level labels, such as buildings and trees. In this research, we developed an algorithm that can perform high-quality and diverse semantic image synthesis using only a small amount of labeled training data. Furthermore, we also developed a method for controlling the layout of generated images without using any labeled training data. We obtained research outcomes containing semantic image synthesis diversification (two domestic meetings and two journals/international conferences), few-shot semantic image synthesis (one domestic meeting and one journal/international conference), and zero-shot control of image generation (one domestic meeting and one journal/international conference).

Academic Significance and Societal Importance of the Research Achievements

本研究成果は、ここ数年で急速に発展している画像生成モデルにおいて、ユーザが介入可能な方法を開拓し、意味ラベルマップなど用いて出力を従来よりも低コストで、柔軟かつ多様に制御可能な方法を示したことに学術的な意義がある。社会的には、コンテンツ産業における創作活動の促進だけでなく、自動運転や医用画像解析の画像認識モデルの精度向上のための訓練データの構築など、本技術の広範な応用が期待できる。

Report

(4 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • Research Products

    (11 results)

All 2022 2021 2020 Other

All Journal Article (4 results) (of which Peer Reviewed: 4 results) Presentation (6 results) (of which Int'l Joint Research: 3 results) Remarks (1 results)

  • [Journal Article] Diversifying detail and appearance in sketch-based face image synthesis2022

    • Author(s)
      Yoshikawa Takato、Endo Yuki、Kanamori Yoshihiro
    • Journal Title

      The Visual Computer

      Volume: 38 Issue: 9-10 Pages: 3121-3133

    • DOI

      10.1007/s00371-022-02538-7

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Controlling StyleGANs using rough scribbles via one‐shot learning2022

    • Author(s)
      Endo Yuki、Kanamori Yoshihiro
    • Journal Title

      Computer Animation and Virtual Worlds

      Volume: 33 Issue: 5

    • DOI

      10.1002/cav.2102

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] User‐Controllable Latent Transformer for StyleGAN Image Layout Editing2022

    • Author(s)
      Endo Y.
    • Journal Title

      Computer Graphics Forum

      Volume: 41 Issue: 7 Pages: 395-406

    • DOI

      10.1111/cgf.14686

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Diversifying Semantic Image Synthesis and Editing via Class‐ and Layer‐wise VAEs2020

    • Author(s)
      Endo Y.、Kanamori Y.
    • Journal Title

      Computer Graphics Forum

      Volume: 39 Issue: 7 Pages: 519-530

    • DOI

      10.1111/cgf.14164

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Presentation] Diversifying detail and appearance in sketch-based face image synthesis2022

    • Author(s)
      Yoshikawa, T., Endo, Y. & Kanamori, Y.
    • Organizer
      Computer Graphics International 2022
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] Controlling StyleGANs using rough scribbles via one-shot learning2022

    • Author(s)
      Yuki Endo, Yoshihiro Kanamori
    • Organizer
      Computer Graphics International 2022
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] User-Controllable Latent Transformer for StyleGAN Image Layout Editing2022

    • Author(s)
      Yuki Endo
    • Organizer
      Pacific Graphics 2022
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] StyleGAN を用いたテキストによる人物画像の服装編集手法2022

    • Author(s)
      吉川 天斗,遠藤 結城, 金森由博
    • Organizer
      第 189 回コンピュータグラフィックスとビジュアル情報学研究発表会
    • Related Report
      2022 Annual Research Report
  • [Presentation] ユーザ制御可能なLatent Transformer を用いたStyleGAN 画像のレイアウト編集2022

    • Author(s)
      遠藤 結城
    • Organizer
      Visual Computing 2022
    • Related Report
      2022 Annual Research Report
  • [Presentation] StyleGAN Prior を用いたFew-shot 意味的画像合成2021

    • Author(s)
      遠藤結城、金森由博
    • Organizer
      Visual Computing 2021
    • Related Report
      2021 Research-status Report
  • [Remarks] 研究プロジェクトページ

    • URL

      http://www.cgg.cs.tsukuba.ac.jp/~endo/projects/clVAE/

    • Related Report
      2020 Research-status Report

URL: 

Published: 2020-04-28   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi