Automatic large-scale classification of bird sounds is strongly improved by unsupervised feature learning.
- Submitting institution
-
Queen Mary University of London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 399
- Type
- D - Journal article
- DOI
-
10.7717/peerj.488
- Title of journal
- PeerJ
- Article number
- ARTN E488
- First page
- e488
- Volume
- 2
- Issue
- 1
- ISSN
- 2167-8359
- Open access status
- Out of scope for open access requirements
- Month of publication
- July
- Year of publication
- 2014
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
1
- Research group(s)
-
-
- Citation count
- 111
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- We introduced a novel unsupervised machine-learning variant for birdsong audio. We went beyond the state of the art in method, recognising hundreds of different bird species (not just a few dozen). This formed the basis of a successful spinout company and app "Warblr Ltd" in 2015, which now has thousands of users all around the UK, received major TV/radio/print news coverage (BBC, Sky News, Reuters, Daily Telegraph...), and has received charity funding (Captain Planet Foundation, $10,000) for a USA launch (currently beta-testing). Led to Stowell invited talks: Universite Paris Sud, and British Trust for Ornithology.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -