A strongly annotated passive acoustic dataset for tropical bird monitoring

TL;DR

This paper introduces PteroSet, a comprehensive, strongly annotated tropical bird sound dataset designed to advance supervised machine learning for biodiversity monitoring in complex ecosystems.

Contribution

The creation and release of PteroSet, a large, richly annotated tropical bird vocalization dataset with a COCO-inspired schema, serving as a benchmark for machine learning research.

Findings

PteroSet contains 563 recordings and 15,372 annotations across 168 species.

The dataset reveals key characteristics of tropical soundscapes, such as acoustic co-occurrence.

A baseline deep learning model for bird detection demonstrates the dataset's practical utility.

Abstract

Passive acoustic monitoring enables continuous, non-invasive biodiversity assessment across diverse ecosystems. The scale of these datasets has driven the adoption of machine learning, with supervised approaches showing strong performance. However, supervised methods require time-resolved annotated datasets, which remain scarce, especially in complex tropical soundscapes. We present PteroSet, a curated dataset of strongly annotated Neotropical bird vocalizations recorded in Puerto Asis (Putumayo) and Pivijay (Magdalena), Colombia, between 2023 and 2025. The dataset comprises 563 recordings (73.62 h) and 15,372 time-frequency annotations, including 6,702 events identified to the species level across 168 species. We release the annotations in a COCO-inspired JSON schema that unifies audio files, taxonomic categories, and labels for machine learning workflows. Beyond providing annotated data, PteroSet serves as a realistic benchmark that highlights key characteristics of tropical soundscapes, including acoustic co-occurrence and domain shift across recording sites. We provide a deep learning baseline for binary bird detection, demonstrating PteroSet's usability and the challenges it presents.