Currently, human-machine communication is far from that of humans. This is due not only to the fact that the ability to transmit information other than the language of machines is still insufficient, but also to the fact that it has not even been clarified what kind of information humans are exchanging with each other in the first place. In our laboratory, what kind of information is exchanged in communication and how is it transmitted? We are also conducting research to exchange that information with computers.
Basic Information
Faculty name/Affiliation
Tomohiro Nagata / Department of Integrated Science and Engineering Computer Science Course
Specialized Fields
Voice information engineering, computer engineering
Research theme
Research on speech synthesis that can express nonverbal information
Research keywords
Nonverbal communication, emotion recognition, speech synthesis, machine learning, social signals
Dialogue speech synthesis Speech is generally considered to be the medium for communicating linguistic messages. However, in the context of communication, we convey not only linguistic messages but also nonverbal messages such as emotions and attitudes. In this project, we are researching speech synthesis techniques for expressing such fire-linguistic messages.
Composition of laughter Humans convey information not only in words, but in various ways. "Laughter" is one of them. Laughter is sometimes referred to as a lubricant in communication. Just as there are many expressions related to laughter such as bitter laughter, ridicule, and ridicule, the information that laughter conveys or conveys is ruthlessly diverse. In this project, we are conducting research to synthesize these various laughter with a computer.
Papers and Conferences Presentation
Paper presentation
Title
Laboratory
Contents
Dimensional paralinguistic information control based on multiple-regression HSMM for spontaneous dialogue speech synthesis with robust parameter estimation
