Where wild things roam: Identifying wildlife with SpeciesNet

2026-03-06 · Source: The latest research from Google · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, AI for Environmental Conservation · Depth: Intermediate, medium

Summary

SpeciesNet, an open-source AI tool developed by Google Research, automatically identifies nearly 2,500 animal categories in camera trap images. Released a year ago, it has been adopted by research groups globally, including Project Lucitania in Colombia and the Idaho Department of Fish and Game, to analyze millions of images. The model was trained on over 65 million labeled images from conservation partners and publicly available repositories, achieving 99.4% accuracy in detecting animals and correctly classifying species 94.5% of the time when identified. SpeciesNet integrates with MegaDetector for animal detection and can process up to 250,000 images daily on a low-end gaming GPU, significantly accelerating wildlife monitoring and conservation efforts. It is also a core component of the Google Cloud-based Wildlife Insights platform.

Key takeaway

For AI Scientists and Conservation Technologists managing large-scale wildlife monitoring projects, integrating SpeciesNet into your workflow can drastically reduce manual image analysis time. Its open-source nature allows for adaptation to local species, and its high processing capacity enables rapid analysis of millions of images, providing timely data for population health assessments and conservation strategies. Consider contributing to its GitHub repository to further refine its capabilities.

Key insights

SpeciesNet is an open-source AI model for automated wildlife identification in camera trap images.

Principles

• Open-sourcing AI tools accelerates conservation research.
• Large, diverse datasets are crucial for robust species classification.
• Automated image analysis enhances wildlife monitoring scalability.

Method

SpeciesNet uses a convolutional neural network, trained on 65M+ images, to classify 2,498 animal categories. It works with MegaDetector to first detect animals, then provides species names and confidence levels.

In practice

• Process 30,000 images/day on a laptop, 250,000+ on a gaming GPU.
• Adapt SpeciesNet by training on local species not in its 2,498 categories.
• Integrate SpeciesNet into existing camera trap data workflows.

Topics

• Wildlife Monitoring
• Species Identification
• Camera Traps
• Deep Learning
• Open-Source AI

Code references

• agentmorris/MegaDetector
• google/cameratrapai

Best for: Computer Vision Engineer, AI Scientist, Research Scientist, AI Researcher, Data Scientist, Domain Expert

Related on AIssential

• How our open-source AI model SpeciesNet is helping to promote wildlife conservation — SpeciesNet is an open-source AI model accelerating wildlife monitoring and conservation through automated image identification.
• How AI is revealing the secret lives of animals from hummingbirds to pumas — AI, GPS, and satellite technologies are profoundly revealing animal behaviors and ecological roles, enhancing conservation efforts.
• AI enables a Who’s Who of brown bears in Alaska — Pose-aware metric learning on stable head features enables individual identification of unmarked animals despite significant physical changes.
• DeepForestVisionV2: Ecology-Driven Taxonomy Expansion for Camera-Trap Monitoring in African Tropical Forests — DeepForestVisionV2 expands camera-trap classification to 64 classes, improving utility across diverse African tropical forest habitats.
• An Open Dataset for the Acoustic Monitoring of Nocturnal Migratory Birds in Europe — Crowd-sourced acoustic data and a specialized deep learning model enhance nocturnal migratory bird monitoring.
• Kiwibit’s AI-powered bird feeder is my new backyard buddy — AI-powered wildlife monitoring systems can enhance nature observation but may exhibit specific algorithmic limitations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The latest research from Google.