Details of Award
NERC Reference : NE/K007270/1
[Environment] Automatic Species Identification
Training Grant Award
- Lead Supervisor:
- Professor N Jones, Imperial College London, Mathematics
- Grant held at:
- Imperial College London, Mathematics
- Science Area:
- Atmospheric
- Earth
- Freshwater
- Marine
- Terrestrial
- Overall Classification:
- Terrestrial
- ENRIs:
- Biodiversity
- Environmental Risks and Hazards
- Global Change
- Natural Resource Management
- Pollution and Waste
- Science Topics:
- None
- Abstract:
- Goal: Using new methods to obtain powerful and generic tools for species identification; towards feature-based fingerprinting. Motivation: A first step of environmental monitoring is establishing the species present. For purposes of automated environmental monitoring we must move towards automated species identification. Automated species identification is a central element of the Natural History Museums Strategy and its future commercial utility. Automation is also a corollary of the NERC 2011 report "Developing a National Strategy in Taxonomy & Systematics". The report emphasises digitisation and the aim of providing appropriate identification tools for all UK organisms: this requires automated species identification to be of maximum utility. Timeliness: the group of Nick Jones (NJ) has recently developed a very powerful tool for probing signal (time-series) data. It is built by integrating thousands of algorithms from across time series analysis in the sciences, and allows an intuitive method for fingerprinting time-series data in which each feature extracted can be interpreted. Norman MacLeod (NM) has recently argued that his expertise in morphometric methods can be applied to spectrograms from signal data. Both NM and NJ have written together on new methods for studying the evolution of function-valued data (like time-series) Workplan: Year 1 (and continuing thereafter): To master and then expand the current tool developed by NJ's group to include methods tailored for audio analysis. To use the signal spectrogram to develop morphometric summaries of signals using NM's tools. Years 1,2: To apply the methods developed to datasets which we already hold. Including ovenbird song records, bat echolocation calls and small samples from the Macaulay Library of animal sounds. For the ovenbirds we have a phylogeny and so will use the comparative method to identify candidate sets of signal features which can serve as fingerprints. Year 1.5 onwards: to find fingerprints (small sets of interpretable signal features) which can be used to distinguish between animal sounds within a large database of diverse animal sounds. Starting first with a wider set of bird sounds and then expanding to other animals. This will be developed from the Macaulay Library.
- NERC Reference:
- NE/K007270/1
- Grant Stage:
- Completed
- Scheme:
- DTG - directed
- Grant Status:
- Closed
- Programme:
- Open CASE
This training grant award has a total value of £75,671
FDAB - Financial Details (Award breakdown by headings)
| Total - Fees | Total - RTSG | Total - Student Stipend |
|---|---|---|
| £13,978 | £5,499 | £56,193 |
If you need further help, please read the user guide.