Nat MethodsJClub
Automated identification of individual birds by song enables multi-year recapture from passive acoustic monitoring data.
Nature Methods · · Journal Article · Open access
Lapp, Lyon + more
Abstract ↗AI summary
The abstract is read at the publisher; the summary is JClub's.
Song-based automated identification of individual male Seiurus aurocapilla from passive acoustic recordings achieves 96% accuracy for individuals not in training data.
- Why it matters: Species-level detections from passive sensors still leave individual tracking dependent on invasive, high-effort capture-mark-recapture, limiting multi-year population studies.
- What they did: A convolutional neural network learns audio features via a location-prediction task, followed by dimensionality reduction and clustering to assign individuals from recordings.
- The result: Applied to 405 male ovenbirds from 126 locations over 4 years, the method enables modeling of abundance and apparent survival, with open-source code for other species.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.
- Open access
- 6 cites