AVISNet: An edge-computing model for acoustic identification of species from the Doñana bird community
4 affiliations
- 1Instituto de Microelectronica de Sevilla (IMSE-CNM), Universidad de Sevilla-CSIC, Seville, Spain
- 2Universidad de Sevilla, Dpto. Ciencias de la Computación e Inteligencia Artificial, Sevilla, Spain
- 3Estacion Biologica de Doñana (EBD), CSIC, Seville, Spain
- 4ICTS-Doñana, Estacion Biologica de Doñana (EBD), CSIC, Seville, Spain
Abstract
The global decline in biodiversity calls for scalable and efficient species monitoring. Birds, as sensitive ecological indicators with rich vocal activity, are particularly well suited for passive acoustic monitoring (PAM), which enables extensive, non-invasive data collection. Yet the large volumes of audio generated by PAM require automated and efficient processing pipelines to extract meaningful biological information in real time. To address this challenge, we integrated automated acoustic species identification directly into embedded devices (i.e., at the edge), enabling autonomous on-site data processing without reliance on cloud-based computation. Our methodology emphasizes model efficiency and leverages lightweight pre-trained architectures (e.g., MobileNet), alongside advanced learning strategies such as knowledge distillation and few-shot learning to overcome data scarcity for under-represented species. This methodology was applied as a case study to acoustic identification of bird species native to Doñana National Park, Spain. We assessed the performance and efficiency of our models at three levels of acoustic data: focal segments, focal recordings, and soundscape recordings. On focal recordings, our models achieve up to 0.70 F1-score in species identification task. At the soundscape level, we evaluated performance using three real-world soundscape datasets, measuring accuracy metrics at both the vocalization and dataset levels. Our lightweight convolutional neural networks achieve high accuracy at the dataset level (F1-score: 0.56–0.72), solid accuracy at the vocalization level (F1-score: 0.29–0.39; mAP: 0.41-0.69), and fast and energy-efficient inference on resource-limited hardware (2.2–4.7× speed-ups vs. BirdNET, 41–71% energy savings). Overall, our results show that these networks achieve functional species recognition performance while substantially reducing inference latency and energy requirements relative to BirdNET, thus enabling on-device deployment on low-power units. The system already scales to 300 + species, with room for further expansion, bridging the gap between general-purpose tools like BirdNET and localized embedded monitoring pipelines.
Related
- Recording context and data splitting can inflate acoustic re-identification accuracy across species and studies
- Automated identification of individual birds by song enables multi-year recapture from passive acoustic monitoring data
- Multicopter-assisted acoustic surveys complement ground-based transects in assessing bat activity and species group composition
- Intelligent identification of dolphin whistle in acoustic signals via deep learning
- Xanthomonas pateli sp. nov. and X. hingoranii sp. nov. reveal a community of non-pathogenic Xanthomonas species complex in citrus
BibTeX
@article{velascomontero2026avisnet,
title = {{AVISNet: An edge-computing model for acoustic identification of species from the Doñana bird community}},
author = {Delia Velasco-Montero and David Solís-Martín and Jorge Fernández-Berni and Giulia Bastianelli and Javier Bustamante},
journal = {PLOS One},
year = {2026},
month = nov,
doi = {10.1371/journal.pone.0360378},
url = {https://doi.org/10.1371/journal.pone.0360378},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - AVISNet: An edge-computing model for acoustic identification of species from the Doñana bird community
AU - Delia Velasco-Montero
AU - David Solís-Martín
AU - Jorge Fernández-Berni
AU - Giulia Bastianelli
AU - Javier Bustamante
PY - 2026
DA - 2026/11/04
JO - PLOS One
DO - 10.1371/journal.pone.0360378
UR - https://doi.org/10.1371/journal.pone.0360378
AB - The global decline in biodiversity calls for scalable and efficient species monitoring. Birds, as sensitive ecological indicators with rich vocal activity, are particularly well suited for passive acoustic monitoring (PAM), which enables extensive, non-invasive data collection. Yet the large volumes of audio generated by PAM require automated and efficient processing pipelines to extract meaningful biological information in real time. To address this challenge, we integrated automated acoustic species identification directly into embedded devices (i.e., at the edge), enabling autonomous on-site data processing without reliance on cloud-based computation. Our methodology emphasizes model efficiency and leverages lightweight pre-trained architectures (e.g., MobileNet), alongside advanced learning strategies such as knowledge distillation and few-shot learning to overcome data scarcity for under-represented species. This methodology was applied as a case study to acoustic identification of bird species native to Doñana National Park, Spain. We assessed the performance and efficiency of our models at three levels of acoustic data: focal segments, focal recordings, and soundscape recordings. On focal recordings, our models achieve up to 0.70 F1-score in species identification task. At the soundscape level, we evaluated performance using three real-world soundscape datasets, measuring accuracy metrics at both the vocalization and dataset levels. Our lightweight convolutional neural networks achieve high accuracy at the dataset level (F1-score: 0.56–0.72), solid accuracy at the vocalization level (F1-score: 0.29–0.39; mAP: 0.41-0.69), and fast and energy-efficient inference on resource-limited hardware (2.2–4.7× speed-ups vs. BirdNET, 41–71% energy savings). Overall, our results show that these networks achieve functional species recognition performance while substantially reducing inference latency and energy requirements relative to BirdNET, thus enabling on-device deployment on low-power units. The system already scales to 300 + species, with room for further expansion, bridging the gap between general-purpose tools like BirdNET and localized embedded monitoring pipelines.
SN - 1932-6203
ER -