An artificial intelligence (AI)-powered application has been developed to identify dinosaur footprints, offering new insights into prehistoric life and potentially earlier bird evolution. The DinoTracker app allows users to upload photos or drawings of footprints via a mobile phone for immediate analysis of their shape and structure.
Dinosaur footprints provide valuable information about how these animals moved and behaved, but their interpretation has historically been challenging due to alterations over time. Traditional methods involved manually built computer databases that linked footprints to specific dinosaurs, a process that experts noted could introduce bias, especially with uncertain or disputed tracks.
To address these issues, a research team led by the Helmholtz-Zentrum research center in Berlin, in collaboration with the University of Edinburgh, created advanced algorithms. The AI system was trained using nearly 2,000 real fossil footprints and millions of simulated examples. These simulations included realistic variations such as compression and edge displacement that occur during fossilization.
The AI model learned to recognize eight key features that differentiate footprints, including toe splay, heel position, ground contact surface area, and weight distribution. After identifying these variations, the system compared new footprints to known fossil examples to determine the most likely dinosaur creator. The algorithm matched classifications made by human experts approximately 90% of the time, even for controversial or difficult-to-identify species.
The AI analysis yielded unexpected connections to birds. For tracks over 200 million years old, the system detected similarities between some dinosaur footprints and the feet of both extinct and modern birds. This finding suggests two possibilities: birds may have emerged tens of millions of years earlier than previously believed, or some early dinosaurs coincidently possessed bird-like feet.
The system also provided new clues regarding mysterious footprints on the Isle of Skye in Scotland. These tracks, formed approximately 170 million years ago, may have been left by some of the oldest known relatives of duck-billed dinosaurs, indicating they are among the earliest examples of this group identified globally.
The new technology aims to expand opportunities for studying dinosaur movement and behavior and enables public participation in fossil research. The study was published in PNAS and received funding from the innovations pool of the BMBF-Project: Data-X, the Helmholtz project ROCK-IT, the Helmholtz-AI project NorMImag, the National Geographic Society, and the Leverhulme Trust.
Dr. Gregor Hartmann of Helmholtz-Zentrum research center stated the method offers an unbiased way to recognize footprint variations and test hypotheses about their makers, making it a tool for research, education, and fieldwork. Professor Steve Brusatte of the University of Edinburgh called the study a significant contribution to paleontology, providing an objective, data-driven approach to classifying footprints, a century-long challenge for experts. He added that the computer network might have identified the world’s oldest birds.
