Chinese researchers develop AI technology to identify individual primates
Chinese researchers have developed an artificial intelligence (AI) system to identify individual primates, offering help in animal research and conservation.
Accurate and quick identification of primates and other species is key to animal behavior and conservation research. A team of researchers from Northwest University, comprising zoology and computer science experts, designed a deep neural network for automated face detection and individual identification of primates and other species.
The AI-system, named Tri-AI, can rapidly detect and identify individual animals from videos and images. It is applicable to multiple species and enables night-time data analysis. Tri-AI has been validated in 41 representative species of primates and four carnivore groups, with an average recognition accuracy of 94.1 percent.
The system can recognize 31 animal images per second, which significantly improves the efficiency of individual data analysis. It can be used as the core component of big data analysis, providing new technical dimensions to zoological research, thus providing reliable technical support for wildlife protection and intelligent management.
The system has obtained three patents. The study was published in the journal iScience.