Google has Open sourcing A AI model, Specieset, which should identify animal species by analyzing photos from camera traps.
Researchers around the world use camera traps – digital cameras associated with infrared sensors to examine wildlife populations. Although these traps can provide valuable knowledge, they create massive amounts of data that take days to weeks to seven.
To help, Google started wildlife Insights about six years ago, an initiative of the Google Earth Outreach Philanthropy program of the company, an initiative of the Google Earth Outreach Philanthropy program. Wildlife Insights offers a platform on which researchers share wildlife pictures online, identify and analyze and work together to accelerate the data analysis of the camera trap.
Many of Wildlife Insights’ analysis tools are powered by Speciesnet, which Google claims to have over 65 million publicly available pictures and pictures of organizations such as the Smithsonian Conservation Biology Institute, the Wildlife Conservation Society, the North Carolina Museum of Natural Sciences and the Zoological Society of London.
According to Google, Spezieset can arrive in one of more than 2,000 labels and animal species, taxa such as “mammals” or “Felidae” and non-animal objects (e.g. “vehicle”).
“The release of Speatatsnet AI model will enable tool developers, academics and biological diversity to scale the monitoring of biological diversity in natural areas,” “Google wrote in a blog post that was published on Monday.
Speciesnet is available on GitHub under an Apache 2.0 license, ie it can be used commercially mainly without restrictions.
It is worth noting that Google is not the only open source tool for automating the analysis of camera trap images. Microsoft’s AI retains good laboratory Pytorch wildlifeA AI framework that offers temporary models that are well coordinated for the detection and classification of animals.
