Takami Yoshida
Papers
10
Total Citations
241
H-Index
6
About
Takami Yoshida is a leading researcher in robot audition and audio-visual integration, whose work has fundamentally advanced how robots perceive and interact with sound in complex environments. His primary research areas include auditory scene analysis, sound source localization, and noise-robust automatic speech recognition (ASR) for mobile robots. Yoshida’s most impactful contribution is the development of a moving microphone array embedded in a quadrocopter for outdoor auditory scene analysis (95 citations), enabling drones to rapidly and widely detect sound sources—a breakthrough with applications in disaster response and search-and-rescue. He also pioneered selectable sound separation on the Texai telepresence system using the HARK robot audition software (38 citations), allowing remote operators to isolate specific speakers in noisy offices. His two-layered audio-visual integration for ASR (35 citations) significantly improved speech recognition robustness against distance and interfering talkers. Additionally, Yoshida addressed practical challenges such as online calibration of asynchronous microphone arrays (30 citations) and ego-motion noise suppression using Semi-Blind Infinite Non-negative Matrix Factorization (21 citations). His innovative fusion of auditory and visual cues, combined with active motion strategies, has set new standards for human-robot communication in real-world settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10Audio-visual speech recognition system for a robot.2 citations · 2010