The IEEE Technical Committee for Audio and Acoustic Signal Processing (AASP) will be launching a set of 'challenges' for specific
problems in the AASP technical scope in order to encourage the devlopment of systems that would have comparable and repeatable
results and at the same time help better defining each of the problems, provide standard datasets for each problem, suggest evaluation metrics
and thus motivate researchers to take part in the whole attempt. For further details visit the
IEEE - AASP challenges webpage.
AASP Challenge - EECS, QMUL
'Detection and Classification of Acoustic Scenes and Events'
Dimitrios Giannoulis, Emmanouil Benetos, Dan Stowell, Mathieu Lagrange,
Mark Plumbley
We have proposed to organise a challenge on the performance evaluation of systems for the detection and classification of acoustic events, which falls into the field of computational auditory scene analysis (CASA) [1]. Humans are able to follow specific sound sources in a complex audio environment with ease and the development of systems that try to mimic this behaviour is an open problem, especially in the case of overlapping sound events.
Currently, there is no active challenge or competition in the CASA field that could encourage research and development towards this direction or provide a source of references of various methodologies and approaches, like the MIREX competition for music information retrieval [2] or the SiSEC evaluation for signal separation [3]. Perhaps the closest is TRECVID Multimedia Event Detection but this is for audio visual event detection [4].
References
[1] D. L. Wang and G. J. Brown (Eds), "Computational auditory scene analysis: Principles, algorithms and applications," IEEE Press/Wiley-Interscience, 2006.[2] MIREX Campaign
[3] SiSEC Evaluation
[4] TRECVID 2011 MED Evaluation track
→ The official web page for the challenge is:
http://www.elec.qmul.ac.uk/digitalmusic/sceneseventschallenge/
→ To obtain a full copy of the Challenge's Proposal click on the pdf link below:
Proposal.pdf
Sound Examples from various Audio Scenes
You can listen to a few examples below to get an idea of how the datasets' sounds would sound:Bus
Office
Sound Example and Annotation for the Office Event Detection Task
You can listen to an example recording of office sounds and see the corresponding annotation:Office Recording
Annotation
IEEE AASP Challenges