Population size estimation
A statistical framework connecting acoustic activity with population-level inference across different vocalization scenarios.
In progressMAASTRICHT UNIVERSITY · BIOACOUSTICS RESEARCH
Listen closer.
Understand more.
We use bioacoustics and machine learning to study animal populations in real-world settings, starting with birds in European zoos.
8Partner zoos across Europe
2,098Individuals
104Bird species
Monthsof acoustic recordings
01 / OUR WORK
Research at the intersection of
sound, ecology, and AI.
BioDCASE 2026
A 2026 bird-counting challenge using real, multi-species zoo aviary recordings. Dataset, baseline, and results are available.
ARIA
A hybrid BirdNET and PERCH framework for recognizing zoo species in overlapping aviary recordings.
Vocalization modeling
Studying temporal patterns, overlapping calls, and interactions in bird vocal activity.
In progressFIELDWORK & DATA
How we listen.
See how aviary recordings become data for species recognition and bird counting.
Explore the recordings and data02 / COMMUNITY
Research happens together.
A record of the challenges and events that brought together people working in bioacoustics, ecology, and machine learning.
03 / OUTPUT
Papers & submissions.
Counting without Seeing: Toward Acoustic Population Estimation from Unsupervised Audio Features
Aysenur Arslan-Dogan · Aki Härmä
ARIA: A Hybrid BirdNET–PERCH Framework for Inventory-Driven Bird Species Recognition in Zoo Aviaries
Emre Argın · Bernardo Amado Costa · Aki Härmä · Aysenur Arslan-Dogan
BioDCASE 2026 Challenge on Bird Population Counting from Passive Acoustic Recordings
Emre Argın · Aysenur Arslan-Dogan · Aki Härmä
BNAIC submission
Title and final publication details will be added after confirmation.
PaCCL: Partition- and Channel-Consistent Call Counting Learning for Passive Acoustic Monitoring of Pied Avocets
Aysenur Arslan-Dogan · Aki Härmä
LET'S CONNECT