Classification of animal sounds in a hyperdiverse rainforest using convolutional neural networks with data augmentation. (December 2022)

• Record Type: Journal Article
• Authors: Sun, Yuren; Midori Maeda, Tatiana; Solís-Lemus, Claudia; Pimentel-Alarcón, Daniel; Buřivalová, Zuzana
• Is Part Of: Ecological indicators. Volume 145(2023)
• Publication Date: 2022-12
• Subjects: Bioacoustics -- Convolutional neural network -- Conservation -- Data augmentation -- Passive 30 acoustic monitoring -- Sound classification -- Tropical forest -- Transfer learning; Environmental monitoring -- Periodicals; Environmental management -- Periodicals; Environmental impact analysis -- Periodicals; Environmental risk assessment -- Periodicals; Sustainable development -- Periodicals; 333.71405
• DOI: 10.1016/j.ecolind.2022.109621
• Languages: English
• ISSNs: 1470-160X

Abstract

Highlights: Machine-learning methods can detect vocalizing animals from rainforest soundscapes. A major obstacle is the necessity of very large training data sets. We overcome this by using data augmentation and transfer learning. Data augmentation and transfer learning lead to >80% accuracy at small sample sizes. Our model can be retrained by undergraduate scientists working with conservation. Abstract: To protect tropical forest biodiversity, we need to be able to detect it reliably, cheaply, and at scale. Automated detection of sound producing animals from passively recorded soundscapes via machine-learning approaches is a promising technique towards this goal, but it is constrained by the necessity of large training data sets. Using soundscapes from a tropical forest in Borneo and a Convolutional Neural Network model (CNN), we investigate i ) the minimum viable training data set size for accurate prediction of call types ('sonotypes'), and ii ) the extent to which data augmentation and transfer learning can overcome the issue of small and imbalanced training data sets. We found that even relatively high sample sizes ( > 80 per sonotype) lead to mediocre accuracy, which however improved significantly with data augmentation and transfer learning, including at extremely small sample sizes (3 per sonotype), regardless of taxonomic group or call characteristics. Neither transfer learning nor data augmentation alone achieved high accuracy. Our results suggest that transfer learning and data augmentation could make the use of CNNs to classify species' vocalizations feasible even for small soundscape-based projects with many rare species. Retraining our open-source model requires only basic programming skills which makes it possible for individual conservation initiatives to match their local context, in order to enable more evidence-informed management of biodiversity.