An AI model designed to identify wildlife is released by Google, called SpeciesNet
Google has introduced an open-source AI model called SpeciesNet, designed to enhance the identification of animal species by analyzing images captured from camera traps. These digital cameras, which are triggered by infrared sensors, are widely used by researchers and conservationists to monitor wildlife populations and study ecological trends. While these camera traps provide invaluable data, they also produce an overwhelming volume of images, requiring significant time and effort to analyze manually. Processing this data can take anywhere from days to weeks, delaying important conservation and research efforts.
To address this challenge, Google launched Wildlife Insights nearly six years ago as part of its Google Earth Outreach philanthropy initiative. This platform allows researchers to share, organize, and analyze wildlife images collaboratively, significantly accelerating the interpretation of camera trap data. Many of the analytical tools integrated into Wildlife Insights rely on SpeciesNet, an advanced AI model that has been trained on an extensive dataset comprising more than 65 million publicly available images. The training data also includes contributions from esteemed 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.
SpeciesNet is capable of classifying images with remarkable accuracy, assigning them to over 2,000 different labels. These classifications range from specific animal species to broader taxonomic groups such as “mammalian” or “Felidae.” Additionally, the model can distinguish non-animal objects, such as vehicles, which may appear in camera trap footage. By making SpeciesNet available as an open-source tool, Google aims to empower developers, academic institutions, conservationists, and biodiversity-focused startups to scale their efforts in monitoring ecosystems and protecting natural habitats.
In a blog post published on Monday, Google emphasized that the release of SpeciesNet under an Apache 2.0 license ensures that researchers and developers can utilize and adapt the model for commercial purposes without major restrictions. This initiative aligns with Google’s broader commitment to leveraging AI for environmental and conservation efforts. However, it is important to note that SpeciesNet is not the only open-source AI solution available for wildlife monitoring. Microsoft’s AI for Good Lab also maintains PyTorch Wildlife, an AI-driven framework that provides pre-trained models specifically optimized for detecting and classifying animals in camera trap images. With more AI-powered tools becoming accessible to researchers and conservationists, the process of studying wildlife populations is set to become more efficient, aiding global efforts to preserve biodiversity.
