Reproduce “Emotional Speech Synthesis for Companion Robot to Imitate Professional Caregiver Speech” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Takeshi Homma, Qinghua Sun, Takuya Fujioka, Ryuta Takawaki, Eriko Ankyu, Kenji Nagamatsu, Daichi Sugawara, Etsuko T. Harada
A speech synthesis method for companion robots automatically extracts and imitates emotional states from professional caregiver speech, aiming to reduce manual emotional parameter tuning. This approach successfully influenced elderly participants' self-perceived arousal, promoting activity in the morning and calmness at night.
Key insights
- Professional caregivers naturally adjust their emotional speech based on context (e.g., 'weaker' voices at night, 'soothing' tones for negative responses), providing a valuable model for robot communication.
- Automatically deriving emotional parameters from target human speech via an emotion recognizer effectively streamlines the emotional speech synthesis process, minimizing manual developer input.
- Emotionally modulated robot speech can exert a specific, measurable influence on human listeners' self-perceived arousal levels, demonstrating potential for therapeutic applications in elder care.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.