Introduction

I am an Assistant Professor at Tampere University, previously known as Tampere University of Technology (until 2018). Prior to this appointment, I was a postdoctoral researcher in the Audio Research Group, and between June 2012-Sept 2014 I worked as postdoctoral researcher at the Department of Signal Processing and Acoustics at Aalto University, Finland.

My PhD thesis on Singing Voice Recognition for Music Information Retrieval was completed at TUT under the supervision of Prof. Tuomas Virtanen and Prof. Anssi Klapuri. My graduate studies were funded by the Graduate School in Electronics, Telecommunications and Automation.

Research

My primary research deals with environmental sounds detection and classification, with the purpose of acoustic scene analysis and understanding. I am interested in the human categorization process as a mechanism for manual annotation of sounds, starting with the cognition aspects involved in the sound recognition, identification of the source, and the semantics of labeling the sound.

The long-term goal of my work is to create intelligent listening machines, that can learn associations between sounds and semantic concepts, using a diverse vocabulary and a large variety of sound examples as their starting point. This involves a combination of signal processing, machine learning, pattern recognition, with added elements of natural language processing, human perception and categorization.

Between 2006-2010, during my graduate studies, I worked in Music Information Retrieval, developing methods for singer identification, lyrics alignment and lyrics transcription from singing (see Singing voice and lyrics recognition section below).

For detailed information on my research activity, see the Research and Publication pages.

DCASE Challenges and Workshops

The Detection and Classification of Acoustic Scenes and Events (DCASE) evaluation campaigns and workshops are an important dissemination channel for our work and a great venue for meeting like-minded researchers. The Audio Research Group has been an active contributor to establishing DCASE as a platform for discussion, comparison and advancement of methods for environmental sound analysis and recognition. With a strong emphasis on reproducible research, DCASE challenges provide open data for research and foster open communication between academics and industry.

Within the DCASE challenges I acted as coordinator of different tasks (acoustic scene classification, sound event detection in real-life audio, rare sound event detection), and contributed to the evaluation framework of the challenge through the evaluation metrics formulation. The outcomes of this work includes many open datasets, an evaluation metric toolbox, and a number of publications that can be found in the publications page.

DCASE Workshop was created on our initiative in 2016, to create a venue for the DCASE Challenge participants and other researchers working on this topic. Since then, the workshop became an independent yearly event, with a different organizing committee each year.

DCASE 2019 Workshop will take place in New York, USA.

For more information on DCASE, see the official DCASE Community webpage.

Open data

Datasets for acoustic scene classification and sound event detection have been published within the DCASE Challenge framework during the years. The datasets are all publicly available, and we strongly support benchmarking novel methods against the different editions of DCASE. for which all results are publicly available.

A list of datasets can be found on the Datasets page:

Teaching

• Academic year 2018-2019: Introduction to Audio Processing
• Academic year 2016-2017: Analysis of Audio, Speech and Music Signals
• Academic year 2015-2016: Analysis of Audio, Speech and Music Signals
• Academic year 2014-2015: Analysis of Audio, Speech and Music Signals
• Academic year 2012-2013: Speech recognition

Personal

In my free time I like to go skating - inline or nordic skating, depending on the season. Most of the year I sneak in a bit of everyday workout by cycling to and from work, and dream of sunny summer days when I can go skating into the sunset. Oh wait, summer sunset in Finland is bit late.

My hobbies also include dressmaking and cooking, with a soft spot for vintage styles and Indian cuisine.