From manual analysis to automation: scaling passive acoustic monitoring for the endangered Saimaa ringed seals

Automated detection methods achieve up to 99.28% accuracy in identifying Saimaa ringed seal vocalizations, enabling scalable acoustic monitoring of this endangered species.

• Why it matters: Manual analysis of large acoustic datasets is time-consuming and limits the ability to monitor elusive pinnipeds effectively, especially for endangered species like the Saimaa ringed seal.
• What they did: Researchers developed automated pulse repetition rate estimation and spectrogram-based neural network detection systems using 12,565 annotated calls from Lake Saimaa, achieving high accuracy in call identification.
• The result: These methods closely match manual analysis, reducing effort and paving the way for long-term, scalable acoustic monitoring to support conservation of the Saimaa ringed seal.