There You Are! Automated Detection of Indris’ Songs on Features Extracted from Passive Acoustic Recordings

Simple Summary Identifying vocalisations of given species from passive acoustic recordings is a common step in bioacoustics. While manual labelling and identification are widespread, this approach is time-consuming, prone to errors, and unsustainable in the long term, given the vast amount of data collected through passive monitoring. We developed an automated classifier based on a convolutional neural network (CNN) for passive acoustic data collected via an in situ monitoring protocol. In particular, we aimed to detect the vocalisations of the only singing lemur, Indri indri. Our network achieved a very high performance (accuracy >90% and recall >80%) in song detection. Our study contributes significantly to the automated wildlife detection research field because it represents a first attempt to combine a CNN and acoustic features based on a third-octave band system for song detection. Moreover, the automated detection provided insights that will improve field data collection and fine-tune conservation practices. The growing concern for the ongoing biodiversity loss drives researchers towards practical and large-scale automated systems to monitor wild animal populations. Primates, with most species threatened by extinction, face substantial risks. We focused on the vocal activity of the indri (Indri indri) recorded in Maromizaha Forest (Madagascar) from 2019 to 2021 via passive acoustics, a method increasingly used for monitoring activities in different environments. We first used indris' songs, loud distinctive vocal sequences, to detect the species' presence. We processed the raw data (66,443 10-min recordings) and extracted acoustic features based on the third-octave band system. We then analysed the features extracted from three datasets, divided according to sampling year, site, and recorder type, with a convolutional neural network that was able to generalise to recording sites and previously unsampled periods via data augmentation and transfer learning. For the three datasets, our network detected the song presence with high accuracy (>90%) and recall (>80%) values. Once provided the model with the time and day of recording, the high-performance values ensured that the classification process could accurately depict both daily and annual habits of indris' singing pattern, critical information to optimise field data collection. Overall, using this easy-to-implement species-specific detection workflow as a preprocessing method allows researchers to reduce the time dedicated to manual classification.