Perch 2.0

Paid

AI for bioacoustics: saving endangered species through sound analysis

4.6

New AI ToolsContact

Type

Saas

Company

Google DeepMind

Visit Website

About Perch 2.0

Perch is an AI model developed by Google DeepMind to analyze bioacoustic data for conservation. The updated Perch 2.0 offers state-of-the-art bird species predictions, improved adaptability to underwater environments like coral reefs, and training on a wider range of animals including mammals, amphibians, and anthropogenic noise. It can disentangle complex acoustic scenes over thousands of hours of audio and answer diverse ecological questions, from tracking individual animals to estimating population abundance. The model is available as an open model on Kaggle, and its vector search capabilities allow scientists to build classifiers from a single example, accelerating the monitoring of endangered species such as Hawaiian honeycreepers and the Plains Wanderer.

Key Features

State-of-the-art off-the-shelf bird species predictions

Improved adaptability to new environments, especially underwater like coral reefs

Trained on nearly twice as much data from public sources (Xeno-Canto, iNaturalist)

Disentangles complex acoustic scenes over thousands or millions of hours

Versatile – answers diverse questions (e.g., birth rates, individual counts)

Vector search enables building classifiers from a single example

Open model available on Kaggle

Integrated into Cornell's BirdNet Analyzer

Pros & Cons

Pros

Finds target sounds up to 50x faster than traditional methods (LOHE Lab study)
Open model free to use on Kaggle
Adapts to new environments and species with minimal training data
Can identify individual animals and track population dynamics
Widely adopted (250,000+ downloads) and integrated into popular tools

Cons

Requires local expert validation to mark search results for classifier training
Performance may depend on quality and diversity of training data
Not explicitly designed for real-time analysis

Best For

Monitoring endangered Hawaiian honeycreepers for avian malaria threatsDiscovering new populations of elusive species like the Plains WandererTracking bird abundance and individual identification without catch-and-releaseAssessing coral reef health through underwater acousticsBuilding classifiers for unique Australian species with BirdLife AustraliaAnalyzing biodiversity in any audio-rich ecosystem

Alternatives to Perch 2.0

PlugSugar

Automate conversations, answer questions with Web Search plugin, and customize ChatGPT experience using powerful AI plugins.

MedARC

Monitor patient progress, streamline processes and generate insights for healthcare providers to improve patient care.

SumUp

Experience Unmatched Sound Quality with AmbienceSoundPro!

Thinking Toolbox

Assess thinking skills, practice critical thinking, and access curated resources to improve cognitive abilities.

Respage

Automate lead acquisition, interact with potential leads, and capture lead information and preferences.

Travel Plan AI

Your personal AI guide for unforgettable journeys.

FAQ

What is Perch?

Perch is an AI model from Google DeepMind that analyzes bioacoustic recordings to help conservationists monitor and protect endangered species by identifying animal vocalizations and patterns.

What new features does Perch 2.0 offer?

Perch 2.0 improves bird species predictions, adapts better to underwater environments, and is trained on a wider range of animals including mammals, amphibians, and anthropogenic noise, using nearly twice as much data as the previous version.

How can I access Perch?

Perch is released as an open model and is available for download on Kaggle.

What are some real-world success stories using Perch?

Perch has been used to monitor Hawaiian honeycreepers (50x faster detection), discover a new population of the Plains Wanderer in Australia, and is integrated into Cornell's BirdNet Analyzer.