Google has opened a species net, an AI model designed to identify animal species by analyzing photos from camera traps.
Researchers around the world use camera traps (digital cameras connected to infrared sensors) to study wildlife populations. However, these traps can provide valuable insights, but generate a large amount of data that takes days to weeks.
Google launched it about six years ago to support Wildlife Insights, an initiative for its Google Earth Autreach Philanthropy program. Wildlife Insights provides a platform for researchers to share, identify and analyze Wildlife images online, and work together to speed up camera trap data analysis.
Many Wildlife Insights analytics tools claim Google is trained on more than 65 million published images and images from organizations such as the Smithsonian Conservation Biology Institute, The Wildlife Conservation Society, the North Carolina Natural Sciences Museum, and the Zoological Society of London.
According to Google, Species Nets can classify images into one of more than 2,000 labels covering animal species, taxa such as “mammals” and “reidae.”
“The release of AI models for species will enable tool developers, academics and biodiversity-related startups to expand their surveillance of biodiversity in natural regions,” Google wrote in a blog post published Monday.
Species NETs are available on GitHub under the Apache 2.0 license. In other words, you can sell the restrictions on the market.
Please note that Google is not the only open source tool for automating the analysis of camera trap images. Microsoft’s AI for Good Lab maintains Pytorch Wildlife, an AI framework that provides pre-trained models that are finely tuned for animal detection and classification.