Salvato in:
Dettagli Bibliografici
Autori principali: Ciapponi, Stefano, Mannini, Leonardo, Scanferla, Jarek, Anderle, Matteo, Farella, Elisabetta
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2509.20103
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912602628554752
author Ciapponi, Stefano
Mannini, Leonardo
Scanferla, Jarek
Anderle, Matteo
Farella, Elisabetta
author_facet Ciapponi, Stefano
Mannini, Leonardo
Scanferla, Jarek
Anderle, Matteo
Farella, Elisabetta
contents This paper introduces WrenNet, an efficient neural network enabling real-time multi-species bird audio classification on low-power microcontrollers for scalable biodiversity monitoring. We propose a semi-learnable spectral feature extractor that adapts to avian vocalizations, outperforming standard mel-scale and fully-learnable alternatives. On an expert-curated 70-species dataset, WrenNet achieves up to 90.8\% accuracy on acoustically distinctive species and 70.1\% on the full task. When deployed on an AudioMoth device ($\leq$1MB RAM), it consumes only 77mJ per inference. Moreover, the proposed model is over 16x more energy-efficient compared to Birdnet when running on a Raspberry Pi 3B+. This work demonstrates the first practical framework for continuous, multi-species acoustic monitoring on low-power edge devices.
format Preprint
id arxiv_https___arxiv_org_abs_2509_20103
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enabling Multi-Species Bird Classification on Low-Power Bioacoustic Loggers
Ciapponi, Stefano
Mannini, Leonardo
Scanferla, Jarek
Anderle, Matteo
Farella, Elisabetta
Sound
Computational Engineering, Finance, and Science
This paper introduces WrenNet, an efficient neural network enabling real-time multi-species bird audio classification on low-power microcontrollers for scalable biodiversity monitoring. We propose a semi-learnable spectral feature extractor that adapts to avian vocalizations, outperforming standard mel-scale and fully-learnable alternatives. On an expert-curated 70-species dataset, WrenNet achieves up to 90.8\% accuracy on acoustically distinctive species and 70.1\% on the full task. When deployed on an AudioMoth device ($\leq$1MB RAM), it consumes only 77mJ per inference. Moreover, the proposed model is over 16x more energy-efficient compared to Birdnet when running on a Raspberry Pi 3B+. This work demonstrates the first practical framework for continuous, multi-species acoustic monitoring on low-power edge devices.
title Enabling Multi-Species Bird Classification on Low-Power Bioacoustic Loggers
topic Sound
Computational Engineering, Finance, and Science
url https://arxiv.org/abs/2509.20103