A workflow for the automated detection and classification of female gibbon calls from long-term acoustic recordings

Abstract

Passive acoustic monitoring (PAM) allows for the study of vocal animals on temporal and spatial scales difficult to achieve using only human observers. Recent improvements in recording technology, data storage, and battery capacity have led to increased use of PAM. One of the main obstacles in implementing wide-scale PAM programs is the lack of open-source programs that efficiently process terabytes of sound recordings and do not require large amounts of training data. Here we describe a workflow for detecting, classifying, and visualizing female Northern grey gibbon calls in Sabah, Malaysia. Our approach detects sound events using band-limited energy summation and does binary classification of these events (gibbon female or not) using machine learning algorithms (support vector machine and random forest). We then applied an unsupervised approach (affinity propagation clustering) to see if we could further differentiate between true and false positives or the number of gibbon females in our dataset. We used this workflow to address three questions: (1) does this automated approach provide reliable estimates of temporal patterns of gibbon calling activity; (2) can unsupervised approaches be applied as a post-processing step to improve the performance of the system; and (3) can unsupervised approaches be used to estimate how many female individuals (or clusters) there are in our study area? We found that performance plateaued with >160 clips of training data for each of our two classes. Using optimized settings, our automated approach achieved a satisfactory performance (F1 score ~ 80%). The unsupervised approach did not effectively differentiate between true and false positives or return clusters that appear to correspond to the number of females in our study area. Our results indicate that more work needs to be done before unsupervised approaches can be reliably used to estimate the number of individual animals occupying an area from PAM data. Future work applying these methods across sites and different gibbon species and comparisons to deep learning approaches will be crucial for future gibbon conservation initiatives across Southeast Asia.

Automated detection of Bornean white-bearded gibbon (Hylobates albibarbis) vocalisations using an open-source framework for deep learning

“…The results show how our best performing model was effective at detecting high-quality H. albibarbis great calls with a low rate of false positive instances. The best performing model (F score: 0.87) exceeded previously reported SVM models for detecting gibbon vocalisations e.g., Hylobates funereus detector, F score: 0.78 (Clink et al 2023). It also performed comparatively to other CNN models, e.g., Nomascus hainanus detector, F score: 0.91 (Dufourq et al 2021), Nomascus concolor detector, F score: 0.92 (Zhou et al 2023).…”

“…Hylobates funereus detector, F score: 0.78 (Clink et al 2023). It also performed comparatively to other CNN models, e.g., Nomascus hainanus detector, F score: 0.91 (Dufourq et al 2021),…”

“…An alternative approach would be an improved post-processing stage to better interpret the output of the model. In audio classification, CNNs inspect audio recordings as image-like segments and so are unable to use broader-scale contextual information, such as whether the current point in a recording is preceded by a target call (Wang et al 2022) In some cases, it may be better to adopt an approach that requires less data to develop, such as an SVM or a GMM (Clink et al 2023).…”

“…Gibbons engage in loud, highly stereotyped song bouts which are largely confined to a specific daily temporal period (Cheyne et al 2008). During a particular calling bout, they usually emit multiple calls, allowing for ample training data (Clink et al 2023). Here, we apply a variation of a pre-defined CNN architecture, DenseNet (Huang et al 2016), to identify female great call vocalisations of the endangered Bornean white-bearded gibbon ( Hylobates albibarbis ) from a long-term acoustic dataset.…”

Pairing a user‐friendly machine‐learning animal sound detector with passive acoustic surveys for occupancy modeling of an endangered primate

“…Other detectors may be more suitable for primate detection than BirdNET, and several are available (e.g., Clink & Klinck, 2019; Dufourq et al, 2021; Heinicke et al, 2015). The potentially substantial acoustic repertoires of some primates may warrant customized signal detection and classification solutions.…”