TinyChirp: Bird Song Recognition Using TinyML Models on Low-power Wireless Acoustic Sensors

Monitoring biodiversity at scale is challenging. De-tecting and identifying species in fine grained taxonomies requires highly accurate machine learning (ML) methods. Training such models requires large high quality data sets. And deploying these models to low power devices requires novel compression techniques and model architectures. While species classification methods have profited from novel data sets and advances in ML methods, in particular neural networks, deploying these state-of-the-art models to low power devices remains difficult. Here we present a comprehensive empirical comparison of various tinyML neural network architectures and compression techniques for species classification. We focus on the example of bird song detection, and more concretely on a data set curated for studying the corn bunting bird species. We publish the data set along with all the code and experiments of this study. In our experiments we comparatively evaluate predictive performance, memory and time complexity of spectrogram-based methods and of more recent approaches operating directly on the raw audio signal. Our results demonstrate that TinyChirp - our approach - can robustly detect individual bird species with precisions over 0.98 and reduce energy consumption compared to state-of-the-art, such that an autonomous recording unit lifetime on a single battery charge is extended from 2 weeks to 8 weeks, almost an entire season.