A changepoint prefilter for sound event detection in long-term bioacoustic recordings

Abstract

Long-term soundscape recordings are useful for a variety of applications, most notably in bioacoustics. However, the processing of such data is currently limited by the ability to efficiently and reliably detect the target sounds, which are often sparse and overshadowed by environmental noise. This paper proposes a sound detector based on changepoint theory applied to a wavelet representation of the sound. In contrast to existing methods, in this framework, theoretical analysis of the detector's performance and optimality for downstream applications can be made. The relevant statistical and algorithmic developments to support these claims are presented. The method is then tested on a real task of detecting two bird species in acoustic surveys. Compared to commonly used alternatives, the proposed method consistently produced a lower false alarm rate and improved the survey efficiency as measured by the precision of the inferred population size. Finally, it is demonstrated how the method can be combined with a simple classifier to detect cat sounds in domestic recordings, which is an example from the Detection and Classification of Acoustic Scenes and Events (DCASE) 2018 workshop. The resulting performance is comparable to the state-of-the-art deep learning models and requires much less training data.

Wind‐robust sound event detection and denoising for bioacoustics

“…Thus, the proposed framework is designed for low training data situations that are common in wildlife research, where recording collection and expert annotation is expensive. In our survey analysis workflow, only the wavelet energy detector needs training, which has a simple structure and so can be trained with less than an hour of data (Juodakis, Marsland, & Priyadarshani, 2021b; Priyadarshani et al, 2020). Neural networks could be used to create noise‐robust detectors that outperform our results if given sufficient data, but this likely means at least thousands of clips, as in, for example, Vickers et al (2021).…”

“…This produces a noise-level estimate, which can be used with standard denoising methods to subtract it and recreate clean sound. For detection, our proposed workflow is to use log-scale spectral subtraction to adjust frequency band energies, and then analyse the adjusted energies using a changepoint detection algorithm (Juodakis, Marsland, & Priyadarshani, 2021b) to detect calls.…”

“…To separate the recorded sound into different frequency bands, we use a WPT representation. Bioacoustic detectors using WPT have been used previously (Juodakis, Marsland, & Priyadarshani, 2021b; Priyadarshani et al, 2020; Zhang & Li, 2015), and a full technical presentation can be found in, for example, Jensen and Cour‐Harbo (2001). Briefly, WPT starts using a pair of filters to decompose the recorded waveform into two series of wavelet coefficients, corresponding to high‐frequency and low‐frequency components.…”

“…The verified detections were used to fit an SCR model, and the density SE estimated by bootstrap (Stevenson et al, 2015) is reported, as well as the coefficient of variation (CoV) to allow differences in the density estimate . We refer the reader to Borchers et al (2015) and Stevenson et al (2015) for a full introduction to acoustic SCR, and to Juodakis, Castro, and Marsland (2021) or Juodakis, Marsland, and Priyadarshani (2021b) for details on formatting data for this type of model.…”

“…In this section, we use data from eight recorders with 6 hr of sound from each, collected in November 2018. These surveys were previously used to evaluate sound detection methods in Juodakis, Marsland, and Priyadarshani (2021b), and further details about the datasets are provided therein.…”