Paper presented by NTT Communication Science Laboratories at ACM Multimedia, the top conference in the multimedia field, was nominated for one of the five Best Paper Award Candidates

NTT Communication Science Laboratories presented a paper at ACM Multimedia (ACM International Conference on Multimedia) 2022, the top conference in the multimedia field, held October 10-14, 2022, at the Lisbon Congress Center in Lisbon, Portugal. This paper was selected from among the oral presentations as a Best Paper Award Candidate [out of 2,472 total submissions, 691 were accepted (acceptance rate: 27.9%), 146 oral presentations (21.1% of the accepted papers), and five Best Paper Award Candidates (0.7% of the accepted papers)].

The paper is as follows:

ConceptBeam: Concept driven target speech extraction

Yasunori Ohishi, Marc Delcroix, Tsubasa Ochiai, Shoko Araki, Daiki Takeuchi, Daisuke Niizumi, Akisato Kimura, Noboru Harada, Kunio Kashino

Using the concepts (semantic content) conveyed by speech and images as a clue, we have developed the first technology in history for extracting speech signals of speakers speaking about a topic or concept of interest in a recording of multiple talkers. This technology can be applied to innovative applications such as listening AIs that focus on a particular topic discussed in a conversation.

NTT Communication Science Laboratories will promote basic research with the aim of building fundamental theories that approach the essence of human beings and information, and creating innovative technologies that will revolutionize society.