Northwestern University in Xi’an, Shaanxi province, made a groundbreaking announcement on May 17 about their open-source AI framework developed by the university's technology team for automated monitoring of wildlife behavior.

Professor Guo Songtao's team at the university devised an AI framework for automatic recognition and measurement based on behavioral analysis principles, leveraging artificial intelligence technology. By integrating "monkey face recognition" technology, they achieved precise automatic identification and measurement of individual wild animals, along with automated monitoring of behavior patterns and time allocation.

Guo Songtao introduces the technological breakthrough.

Guo (first from the left) and his team work in the laboratory.

Animal behavior, a complex yet common natural phenomenon, has been a subject of study for over a century, drawing insights from various fields such as evolutionary biology, ecology, sensory science, and neurobiology. Accurate observation, recording, and quantitative analysis of animal behavior remain fundamental to its research, enabling a deeper understanding of behavioral patterns.

While traditional methods relied on visual observation, recent advancements have shifted towards video recordings for more precise behavioral analysis. However, manual review of video data for even a few days can consume hundreds of hours, limiting comprehensive insights into animal behavior patterns.

The AI framework developed by Professor Guo Songtao's team has effectively addressed these challenges, offering a solution for automatic behavior recognition and measurement using artificial intelligence technology.

To facilitate rapid technological advancements and provide essential data references for future animal behavior research, the team has open-sourced the framework and published it as the "Artificial Intelligence Framework for Automatic Quantitative Measurement of Mammalian Behavior" in the prestigious journal "Integrative Zoology" under the Chinese Academy of Sciences.

Professor Guo Songtao highlighted the successful validation of the framework across various species, emphasizing its potential applications in monitoring captive animals' daily activities, analyzing wild animal behavior, studying habits, assessing experimental animals' health and welfare, and even analyzing human behavior and health characteristics.