If you look at a photo of leopards, would you be able to tell which two are related based on their spots?
Unless you’re a leopard expert, the answer probably isn’t, says Tanya Berger-Wolf, director of the Translational Data Analytics Institute (TDAI) at Ohio State University. But, she says, computers can.
Berger-Wolf and her team are pioneering a new field called imageomics. As the name suggests, imageomics uses machine learning to extract biological data from photos and videos of living organisms. Berger-Wolf and her team recently began collaborating with researchers studying leopards in India to compare maternal and infant spotting patterns using algorithms.
“Images have now become the most abundant source of information, and we have the technology. We have computer vision machine learning,” says Berger-Wolf. She likens this technology to the invention of the microscope, which allows scientists to look at wildlife in a completely different way.
Building on TDAI’s open source platform called game book, which allows naturalists to collect and analyze images, the team is now turning to generative AI approaches. These programs use existing content to generate meaningful data. In this case, they’re trying to analyze crowdsourced images to compute biological features that humans might naturally overlook, such as the curvature of a fish’s fin or a leopard’s spots. The algorithms scan images of leopards available online, from social media to digitized museum collections.
Simply put, the algorithms “quantify the similarity,” she says. The goal is to help naturalists solve a data shortage problem and, ultimately, better protect animals at risk of extinction.
Ecologists and other naturalists are currently facing a data crisis – it is tedious, expensive and time consuming for people to spend time in the field guarding animals. Because of these challenges, 20,054 species on the International Union for Conservation of Nature (IUCN) Red List of Endangered Species are labeled as “data defective,” meaning there is not enough information to make a good estimate of the risk of extinction. As Berger-Wolf sums it up, “biologists make decisions without having good data about what we’re losing and how quickly.”
The platform started supervised learning – Berger-Wolf says the computer uses algorithms that are “simpler than Siri” to count how many animals are in the image, as well as where it was taken and when, which could contribute to statistics such as the number of populations. AI can do this not only at a much lower cost than hiring humans, but at a faster pace. In August 2021, the platform automatically analyzed 17 million images.
There are also barriers that only a computer can seem to overcome. “People aren’t the best at figuring out what the informational aspect is,” she says, noting how people are biased in how we view nature, focusing primarily on facial features. Instead, AI can scan for features humans would likely miss, such as the color range of a tiger moth’s wings. A March 2022 study found that the human eye couldn’t distinguish male polymorphic wood tiger moth genotypes — but moth vision models with ultraviolet light sensitivity could.
“That’s where all the real innovation is in all of this,” says Berger-Wolf. The team implements algorithms that create pixel values from patterned animals, such as leopards, zebras and whale sharks, and analyze hot spots where the pixel values change the most — it’s like comparing fingerprints. With these fingerprints, researchers can track animals non-invasively and without GPS collars, count them to estimate population sizes, understand migration patterns and more.
As Berger-Wolf points out, population size is the most basic measure of a species’ well-being. The platform scanned 11,000 images of whale sharks to create hotspots and help researchers identify individual whale sharks and track their movement, leading to updated information about their population sizes. These new data prompted the IUCN to change the whale shark’s conservation status from “vulnerable” to “endangered” in 2016.
There are also algorithms that use facial recognition for primates and cats, which have been shown to be about 90 percent accuratecompared to people who are about 42 percent accurate.
Generative AI is still a fast-growing field when it comes to conservation, but Berger-Wolf is hopeful. For now, the team is cleaning up preliminary data from leopard hot spots to ensure the results aren’t data artifacts — or flawed — and are genuine biologically meaningful information. If meaningful, the data can teach researchers how species respond to changing habitats and climates and show us where humans can help.
SOURCE – www.theverge.com