Dinosaur footprints, often among the most common types of dinosaur fossils discovered, present a unique challenge to paleontologists. Sometimes, researchers find a single, isolated footprint, while other times they encounter scattered, irregular tracks resembling a dinosaur’s dance across the landscape. Determining which dinosaur made which footprint, however, has remained a difficult task – until now.
Researchers have developed a new methodology utilizing artificial intelligence to help identify the dinosaur responsible for each footprint, based on eight distinct characteristics of the tracks. This development, announced on , promises a more objective approach to paleontological analysis.
“What we have is important because it provides an objective way to classify and compare tracks, reducing reliance on subjective human interpretation,” said Dr. Gregor Hartmann, a physicist at the Helmholtz-Zentrum in Germany and a lead researcher on the project. The research was published in the journal Proceedings of the National Academy of Sciences.
The new method was honed through an analysis of 1,974 footprint silhouettes spanning 150 million years of dinosaur history. The AI was able to discern eight key features that explain variation in footprint shapes, including toe spread, heel position, the size of the contact area, weight distribution and the degree of toe attachment.
The challenge of footprint identification, as explained by University of Edinburgh paleontologist Professor Steve Brusatte, lies in the complexity of the process. “When we find a dinosaur footprint, we try to do the Cinderella thing and find the foot that matches the slipper,” Brusatte said. “But it’s not so simple, because the shape of a dinosaur footprint depends not only on the shape of the dinosaur’s foot but also the type of sand or mud it was walking through, and the motion of its foot.”
Previous AI systems relied on footprints that had already been labeled by paleontologists. However, the team recognized a potential flaw in this approach. “You never find a footprint and alongside [it] the dinosaur that had made this footprint,” Hartmann explained. “So, no offense to paleontologists and such, but most likely some of these labels are wrong.”
To address this, the researchers took a different approach, feeding their AI system with 2,000 unlabeled footprint silhouettes. The system then analyzed the imprints, determining their similarities and differences based on the features it identified as meaningful. This allowed the AI to establish its own classification system, independent of potentially inaccurate pre-existing labels.
The resulting app, dubbed DinoTracker, allows researchers and dinosaur enthusiasts to upload a photo or drawing of a footprint via a mobile phone and receive an immediate analysis. The app estimates which type of dinosaur likely made the track by evaluating its shape and structure, with accuracy rivaling that of human experts – achieving approximately 90% classification accuracy.
Beyond simply identifying dinosaurs, the AI has also yielded unexpected insights into the origins of birds. Analysis of seven small, three-toed footprints, approximately 210 million years old from South Africa, supported previous assessments that these tracks closely resemble those of birds. These footprints are 60 million years older than the oldest known bird fossils, potentially pushing back the origin of birds further into prehistory.
“This doesn’t prove they’re from birds, of course,” Brusatte clarified, suggesting the tracks could have been made by previously unknown dinosaur ancestors of birds, or even dinosaurs unrelated to birds that simply possessed bird-like feet. “So we need to take this seriously and find an explanation for it.”
The variability in footprint size also presents a challenge. Footprints can range dramatically in size, from those of small, chicken-sized carnivorous dinosaurs to those of long-necked dinosaurs the size of a bathtub, according to Brusatte. He noted that finding a fossilized skeleton alongside a trackway is a rare occurrence, emphasizing the detective work involved in identifying the dinosaur responsible for a given set of footprints.
DinoTracker represents a significant advancement in paleontological research, offering a more objective and efficient method for analyzing fossil tracks. By leveraging the power of AI, researchers are not only gaining a better understanding of dinosaur behavior and evolution but also uncovering potential clues about the origins of birds, hidden in plain sight within the ancient record.
