A new AI-driven tracking tool called replicAnt holds promise for advancing biology research on animal behavior, offering biologists a faster and less labor-intensive way to collect and process data. The tool simplifies and streamlines the creation of training images for machine learning tools, automating the annotation process and incorporating variations in species and environments. By generating thousands of annotated images with a click of a mouse, replicAnt speeds up data collection and analysis, potentially increasing the speed and robustness of using AI tools in animal research. The work, led by Fabian Plum, a PhD researcher at Imperial College London’s Department of Bioengineering, has been published in Nature Communications.
replicAnt lowers the entry barrier for biologists to use machine learning to optimize their work, said Fabian Plum. Understanding animal behavior, particularly as our climate changes, is crucial. We hope our tool can help to make the time-intensive process of collecting crucial data easier and faster.
The new tracking tool builds on the researchers’ previous work with scAnt, a 3D scanner that creates high-resolution 3D models of small animals. The 3D models generated by scAnt are then used in replicAnt, which leverages the 3D software Unreal Engine, to produce training images for detecting and tracking animals both in the lab and in nature. This approach significantly reduces the time and effort required by researchers, freeing up their resources and streamlining their work.
To showcase the effectiveness of replicAnt, the researchers trained neural networks using the generated images. These neural networks, which consist of algorithms that recognize underlying relationships in data, were able to identify individuals and track their movements across different environments without the need for extensive hand-labeling of real images.
The project was funded by Imperial College London’s President’s PhD Scholarship awarded to Fabian Plum and the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program, led by Dr. David Labonte at the Department of Bioengineering.
In addition to enhancing biology research, the researchers believe that replicAnt’s real-time movement data can have further applications in fields such as character movement in films and video games. This illustrates the impact that AI-driven tools can have in various industries beyond scientific research.
With its ability to expedite data collection and analysis, replicAnt holds great potential for advancing the understanding of animal behavior, especially in the face of a changing climate. By making the process of collecting crucial data easier and faster, this tool equips biologists with a valuable resource to accelerate their work and drive new discoveries in the field of biology.
