Multi-Input Deep Learning and Its Application to Video Recognition

• Project/Area Number: 15K16019
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Perceptual information processing
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: Inoue Nakamasa 東京工業大学, 情報理工学院, 助教 (10733397)
• Project Period (FY): 2015-04-01 – 2019-03-31
• Project Status: Completed (Fiscal Year 2018)
• Budget Amount: ¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000); Fiscal Year 2018: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000); Fiscal Year 2017: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000); Fiscal Year 2016: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000); Fiscal Year 2015: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
• Keywords: 深層学習 / 映像認識 / パターン認識

Outline of Final Research Achievements

In this project, we proposed a deep learning method for video recognition. The proposed method is based on vocabulary expansion using word vectors. Its performance is demonstrated on the TRECVID video dataset. We presented this work at ACM Multimedia.

Academic Significance and Societal Importance of the Research Achievements

本研究の成果は、映像や画像を認識するための人工知能技術に関するものである。画像データとテキストデータの情報を組み合わせることで、認識精度が向上することを示した。これは映像のどの部分に何があるかを詳細に検索する次世代の検索システムに役立つ技術である。

Research Products

• [Journal Article] [Invited Paper] Semantic Indexing for Large-Scale Video Retrieval (2016)
  • Author(s): Nakamasa Inoue, Koichi Shinoda
  • Journal Title: ITE Transactions on Media Technology and Applications Volume: 4 Issue: 3 Pages: 209-217
  • DOI: 10.3169/mta.4.209
  • NAID: 130005161897
  • ISSN: 2186-7364
  • Peer Reviewed / Open Access
• [Journal Article] Fast Coding of Feature Vectors using Neighbor-To-Neighbor Search (2015)
  • Author(s): Nakamasa Inoue, Koichi Shinoda
  • Journal Title: IEEE Transactions on Pattern Analysis and Machine Intelligence Volume: 99 Issue: 6 Pages: 1-16
  • DOI: 10.1109/tpami.2015.2481390
  • NAID: 120006582446
  • Peer Reviewed / Acknowledgement Compliant
• [Presentation] Activity Detection in Extended Video using Action Tubelets (VANT at TRECVID 2018) (2018)
  • Author(s): Nakamasa Inoue, Chihiro Shiraishi, Aleksandr Drozd, Koichi Shinoda, Shi-wook Lee, Alex Chichung Kot
  • Organizer: TRECVID
  • Int'l Joint Research
• [Presentation] 画像・映像認識 (2018)
  • Author(s): 井上 中順
  • Organizer: 人工知能学会
  • Invited
• [Presentation] 単語ベクトルによる語彙拡張を用いた映像のセマンティックインデクシング (2015)
  • Author(s): 井上 中順, 篠田 浩一
  • Organizer: 電子情報通信学会 PRMU研究会
  • Place of Presentation: 信州大学
  • Year and Date: 2015-12-21
• [Presentation] Vocabulary Expansion Using Word Vectors for Video Semantic Indexing (2015)
  • Author(s): Nakamasa Inoue, Koichi Shinoda
  • Organizer: ACM Multimedia
  • Place of Presentation: Brisbane, AUS
  • Year and Date: 2015-10-26
  • Int'l Joint Research