Academic Biography
Toni Heittola is Machine Learning Lead at Resoniks and a Visiting Researcher at Tampere University, where he previously held a postdoctoral research fellowship affiliated with the Tampere Institute for Advanced Study. He earned his Ph.D. in 2021 from Tampere University, Finland, and was an active member of the university's Audio Research Group from the early 2000s. His early research focused on musical genre and instrument classification, while his doctoral work advanced computational auditory scene analysis, particularly multi-source sound event detection.
Dr. Heittola’s research lies at the intersection of machine learning and audio signal processing, with a focus on environmental audio analysis, sound event detection, and acoustic scene classification. He has contributed to a wide range of academic and industrial projects, including several EU- and nationally funded initiatives such as MARVEL, EVERYSOUND, and SmartSound, where he developed scalable audio recognition systems for smart cities, healthcare, and assistive technologies.
He has authored over 60 peer-reviewed publications, including 11 journal articles, 40 conference papers, and two book chapters, cited over 8000 times on Google Scholar (h-index: 35) and over 3400 times on Scopus (h-index: 22). He is a prominent member of the Detection and Classification of Acoustic Scenes and Events (DCASE) community, where he has served as task coordinator for many DCASE Challenge tasks — designing evaluation protocols, building public datasets, developing baseline systems, and creating reproducibility frameworks and metrics — and as publication chair for several editions of the DCASE Workshop.
A strong advocate for open science, Dr. Heittola has released more than ten publicly available datasets and open-source tools — including sed_eval, dcase_util, and sed_vis — supporting benchmarking and reproducible research. He also maintains the DCASE community website and has built a suite of utilities for collaborative research and community engagement, helping grow DCASE's global participation and academic impact over the years.
In 2019, he co-presented the DCASE tutorial at the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), and in 2021, co-authored the widely cited tutorial article "Sound Event Detection: A Tutorial" in the IEEE Signal Processing Magazine. His work has been recognized with multiple honors, including the IEEE Best Paper Award (2023) and a Doctoral Dissertation Award (2022). At Resoniks, he now applies this research background to AI-powered acoustic testing for manufacturing quality control.