Digital Newsletter
Each week our editor Phil Alsop rounds up the most popular articles, videos and expert opinions. We compile this into a Digital Newsletter and send it straight to your inbox every week.
Digital Magazines
We'll let you know each time a new edition of Digitalisation World is released so that you're always kept up-to-date with the latest and greatest news and press releases.
Video Magazines
The Digitalisation World Video magazine contains the latest Zoom interviews with experts in the industry.
The SA-FARI system is presented as an AI tool intended to support wildlife research projects. It has been developed through collaboration within an international consortium, with contributions from the University of Bristol. The system is designed to autonomously detect, identify, and track individual animals in video footage.
The SA-FARI project, led by ConservationX Labs (CXL) in collaboration with META, builds on the Segment Anything Model 3 (SAM3). It uses vision-language techniques to identify, segment, and track objects by combining text-based and visual inputs.
A key feature of the approach involves the use of “masklets,” which allow researchers to track animals by outlining them and separating them from the background. This approach is intended to reduce the amount of manual analysis required for camera trap footage and to support faster and more consistent extraction of data for behavioural and population studies.
The associated research paper has been accepted for presentation at the Conference on Computer Vision and Pattern Recognition (CVPR). The project is in an early stage of development and has received attention within the computer vision research community.
Professor Tilo Burghardt from the University of Bristol has contributed to the project in the area of animal biometrics, reflecting ongoing research at the university related to ecological applications of AI.
The project involved participation from several research institutions, including the Max Planck Institute for Evolutionary Anthropology and the Senckenberg Museum of Natural History. Together, they worked with a dataset of more than 11,000 wildlife videos to support training and evaluation of the system for detecting and tracking multiple animal species at a fine-grained level.
The project also provides a dataset that is intended to be accessible to researchers. Future development directions include expanding capabilities such as animal pose tracking and generating natural language descriptions of observed behaviour.
