JST CREST project on Smart Seminar Room based on Multi-Modal Recognition of Verbal and Non-Verbal Information

• CREST Project Overview
  Based on multi-modal analysis and modeling of speech communication in human activities, we design IT-enhanced environments for seminars and meetings, which provide real-time supports to participants and generate effective archive for later reference. In addition to the conventional approach to processing of main speaker's utterances, a novel indexing approach focusing on interaction with the audience is investigated.

• Symposium after ICASSP 2012
  Date: April 1-2, 2012
  Venue: Kyoto University Clock Tower Centennial Hall, Kyoto, Japan

• Major Publications and Invited Talks
  • Multi-modal sensing and analysis of poster conversations with smart posterboard.
    APSIPA Trans. Signal & Information Processing, Vol.5, No.e2, pp.1--12, 2016.
  • Smart posterboard: Multi-modal sensing and analysis of poster conversations.
    (APSIPA ASC 2013 plenary overview talk) Paper PDF | Slide PDF
  • Multi-modal Sensing and Analysis of Poster Conversations: Toward Smart Posterboard
    (SIGDIAL 2012 keynote speech) Paper PDF | Slide PDF
  • Demonstration at ICASSP 2012

    (Followings are in Japanese)

  • スマートポスターボード: ポスター会話のマルチモーダルなセンシングと解析
    (人工知能学会 ＡＩチャレンジ研究会 招待講演; Challenge-B303-1 (2014))
  • スマートポスターボード: ポスター発表における場のマルチモーダルなセンシングと認識
    (電子情報通信学会 パターン認識・メディア理解研究会 招待講演; PRMU2012-167 (2013))
  • スマートポスターボード: ポスター会話のマルチモーダルなセンシングと認識
    (電子情報通信学会 音声研究会 招待講演; SP2012-51 (2012))