Yoshiyuki Toyoda

Papers

Total Citations

H-Index

About

Yoshiyuki Toyoda is a pioneering researcher in the field of robotic audition and environmental sound recognition. His work focuses on enabling machines—particularly robots and intelligent computer systems—to interpret and respond to non-speech sounds in real-world environments. Toyoda’s major contributions include the development of neural network-based systems that recognize environmental sounds using innovative input representations, such as time-frequency intersection patterns and combinations of instantaneous power and spectral features. His most cited paper, “Environmental sound recognition by multilayered neural networks” (2004), has garnered 26 citations and addresses the challenge of creating robust recognition systems without exhaustive sound databases. Subsequent works, including his 2012 study on time-frequency intersection patterns (11 citations), further advanced the field by improving recognition accuracy through multi-stage perceptron architectures. Toyoda’s research is notable for its practical emphasis on real-world applicability, tackling the inherent variability and unpredictability of environmental sounds. His work has laid foundational groundwork for applications in robotics, smart environments, and human-computer interaction, demonstrating how neural networks can bridge the gap between controlled laboratory conditions and dynamic, noisy real-world settings.