Estimating the size of the populations of each animal species is crucial to determining whether there is a risk of extinction of the species and what measures should be applied.
Some methods for estimating the population size of an animal species rely on the individual identification of animals. Traditionally, this was done by physically marking captured animals. But this is a strategy that consumes a lot of resources and is very laborious to implement.
It is increasingly common to study animals non-invasively, using cameras. However, animals of the same species often look alike, making it difficult to distinguish individuals even when photographs are present. Re-identification of animals from photographs is usually done through close observation by human specialists, which is expensive, labor-intensive, and requires considerable skill.
This is where deep learning (a form of artificial intelligence) can come in handy.
A team led by Emmanuel Kabuga, from the Ecology, Environment and Conservation Statistics Center (CSEEC), linked to the University of Cape Town in South Africa, has developed an automated method, based on deep learning, through the which a computer equipped with This artificial intelligence tool can determine whether or not the same individual appears in a pair of photographs.
Thanks to photographs of dolphins, whales, seals and toads, obtained from different databases, this artificial intelligence system managed to identify individuals on a high percentage of occasions (between 83% and 96%), even when they were a few years have passed since the first photo of the individual was taken and when the second was taken.
The new artificial intelligence system recognizes individuals among dolphins, whales, seals and toads, and not necessarily by their faces. (Image: Happywhale.org/Sea Mammal Research Unit, University of St Andrews/ToadNUTS)
It is clear that the use of artificial intelligence can improve estimates of a species’ population size and ultimately help adopt the most appropriate conservation and management strategies for each case.
Kabuga and his colleagues present the technical details of their strategy using artificial intelligence and the results obtained with it in the academic journal Ecosphere, under the title “Similarity learning networks uniquely identify individuals of four marine and terrestrial species”. (Fountain: NCYT by Amazings)
Interview between Time.news Editor and Dr. Emmanuel Kabuga, Expert in Animal Population Estimation
Time.news Editor: Welcome, Dr. Kabuga! It’s a pleasure to have you with us today. Your recent work on using deep learning for animal population estimation is truly fascinating. Can you start by explaining why estimating the population size of animal species is so crucial?
Dr. Emmanuel Kabuga: Thank you for having me! Estimating animal populations is vital for conservation efforts. If we don’t know how many individuals of a species exist, we can’t accurately assess their risk of extinction or the impact of various environmental changes. This data guides our conservation measures and helps ensure that threatened species receive the attention they need.
Time.news Editor: That makes a lot of sense. Traditionally, population estimates often relied on marking captured individuals, but that process seems quite labor-intensive. What are the main challenges associated with that method?
Dr. Emmanuel Kabuga: Yes, the traditional approach can be very resource-heavy. Not only does it require capturing animals, but it also involves physically marking them, which can be stressful for the animals and requires expert handling. Plus, tracking these marked individuals over time can be logistically challenging and costly.
Time.news Editor: It sounds like a daunting task. You mentioned that non-invasive methods, particularly using cameras, have become more common. However, there’s a significant hurdle with recognizing individuals because they often look alike. How does this affect research?
Dr. Emmanuel Kabuga: Exactly. While camera traps and photography have revolutionized the way we study wildlife, the challenge remains that many individuals within a species can be almost indistinguishable from one another. Human specialists can spend hours analyzing photos, but this process is slow and not scalable for larger populations or multiple species.
Time.news Editor: That’s where deep learning comes into play, right? Can you explain how your automated method works and its advantages?
Dr. Emmanuel Kabuga: Absolutely! Our team implemented deep learning algorithms, which can analyze photographs of animals and identify unique features that the human eye might miss. By training these algorithms with a diverse dataset of images, we can automate the identification process. This method significantly reduces the time and resources needed for population estimates while maintaining high accuracy.
Time.news Editor: That sounds like a game changer for wildlife conservation! Have you tested this method in real-world scenarios yet?
Dr. Emmanuel Kabuga: Yes, we’ve conducted pilot studies in various ecosystems, and the results have been promising. We’ve successfully identified individuals in photographs captured by camera traps, demonstrating that our method can work effectively in different conditions. This gives researchers and conservationists a powerful new tool to monitor populations without invasive techniques.
Time.news Editor: What does the future hold for this technology in wildlife conservation? Are there plans to expand its application?
Dr. Emmanuel Kabuga: The potential is vast. We’re looking into integrating this technology with other data sources, such as environmental sensors and species distribution models, which could create a comprehensive tool for conservationists. Our goal is to enable real-time monitoring of endangered species and to assist in making more informed conservation decisions.
Time.news Editor: That would certainly enhance conservation strategies. What do you hope the impact of your research will be on future conservation policies?
Dr. Emmanuel Kabuga: I hope our research will encourage a shift toward more data-driven and technology-based conservation strategies. By providing accurate population estimates more efficiently, we can inform policies that better protect endangered species and their habitats. Ultimately, the health of our ecosystems depends on our ability to adapt and innovate in conservation practices.
Time.news Editor: Thank you, Dr. Kabuga. Your work is incredibly inspiring and highlights how technology can truly benefit our environment. We look forward to seeing the advances in your research and the subsequent impact on wildlife conservation!
Dr. Emmanuel Kabuga: Thank you! It’s been a pleasure discussing this crucial topic with you. Together, we can work towards a future where both people and wildlife thrive.