Illuminant Estimation Under Multiple Colored Lighting Using Color Lines

• Project/Area Number: 18K18080
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 61010:Perceptual information processing-related
• Research Institution: Osaka Institute of Technology
• Principal Investigator: Jinno Takao 大阪工業大学, 情報科学部, 講師 (10636070)
• Project Period (FY): 2018-04-01 – 2022-03-31
• Project Status: Completed (Fiscal Year 2021)
• Budget Amount: ¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000); Fiscal Year 2020: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000); Fiscal Year 2019: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000); Fiscal Year 2018: ¥2,860,000 (Direct Cost: ¥2,200,000、Indirect Cost: ¥660,000)
• Keywords: カラーライン / 照明色推定 / 照明光分離 / 有彩色照明 / 可視化 / X線画像 / 非破壊検査 / CNN / 固有画像 / ColorLine / 照明推定 / color line / 領域抽出 / ノイズ低減

Outline of Final Research Achievements

We have proposed and demonstrated the effectiveness of a framework for estimating illumination color using color lines, estimating illuminated color maps using the estimated illumination color, and intrinsic image decomposition. The results of our research will make a significant contribution to the fields of computer vision and computer graphics.
We have proposed A visualization system effective for nondestructive inspection for faint X-ray images. The system allows the user to select the target of visualization, and thus can capture faint or unknown features that are difficult to detect. The results of this research will a significant contribute to nondestructive inspection and medical imaging using X-rays.

Academic Significance and Societal Importance of the Research Achievements

複数の有彩色照明下のシーンに対し，単一の画像のみで色恒常性再現を実現する手法は，ほとんど提案されていない．本研究成果は，コンピュータビジョンのロバスト性向上に資する内容であり，社会的意義が大きい．また，これまでの灰色仮説に基づくアプローチと全く異なるカラーラインを用いた推定手法は，既存手法との組み合わせによる精度向上も望め，学術的にも意義がある．
X線を用いた非破壊検査は日常的な点検作業に有用であり，その作業時間短縮および精度向上可能な本研究成果は大きな社会的意義を持つ．また，ユーザが選択的に可視化対象を選択できることで，対象を限定せず未知の特徴の検出にも利用でき，学術的な意義も大きい．

Research Products

• [Journal Article] Estimation of Multiple Illuminant Colors Using Color Line Features (2022)
  • Author(s): Quan XIU HO, Takao JINNO, Yusuke UCHIMI, and Shigeru KURIYAMA
  • Journal Title: IEICE Transactions on Information and Systems
  • Peer Reviewed
• [Journal Article] A Novel Approach for Investigating Upper Airway Hyperresponsiveness Using Micro-CT in Eosinophilic Upper Airway Inflammation such as Allergic Rhinitis Model (2019)
  • Author(s): Dan Van Bui, Akira Kanda, Yoshiki Kobayashi, Yoshiko Sakata, Yumiko Kono, Yoshiyuki Kamakura, Takao Jinno, Yasutaka Yun, Kensuke Suzuki, Shunsuke Sawada, Mikiya Asako, Akihiko Nakamura, David Dombrowicz, Keita Utsunomiya, Tanigawa Noboru, Koichi Tomoda and Hiroshi Iwai
  • Journal Title: Biomolecules Volume: 9(7) Issue: 7 Pages: 252-263
  • DOI: 10.3390/biom9070252
  • Peer Reviewed / Open Access / Int'l Joint Research
• [Presentation] STUDY ON INTRINSIC IMAGE WITH SINGLE IMAGE (2022)
  • Author(s): Kakeru HARA, Hiromitsu ISOBE, Takao JINNO
  • Organizer: EUSIPCO2022
  • Int'l Joint Research
• [Presentation] Learning-based Color Estimation of Lighting with Color Lines (2020)
  • Author(s): Quan Xiu Ho, Yusuke Uchimi, Takao Jinno, Shigeru Kuriyama
  • Organizer: International Conference on Advanced Informatics: Concept Theory and Applications (ICAICTA 2020)
  • Int'l Joint Research
• [Presentation] Selective visualization for X-ray Inspection of Power Transmission Cable Joints (2020)
  • Author(s): Yuta Suzuki, Kakeru Hara, Takao Jinno, Naohiro Hozumi, Yoshihisa Nagoya, Kimihiro Iwasaki
  • Organizer: CIGRE-AORC Technical Meeting 2020
  • Int'l Joint Research
• [Presentation] Evaluation of Estimation Method for Multiple Illuminant Colors using Color Lines (2018)
  • Author(s): Quan Xiu Ho, Yusuke Uchimi, Takao Jinno and Shigeru Kuriyama
  • Organizer: 5th International Conference on Advanced Informatics, Concepts, Theory, and Applications (ICAICTA)
  • Int'l Joint Research