In 1956, a 23-year-old Canadian scientist named Dr. Anne Innis Dagg did one thing extraordinary: she traveled alone to South Africa to review giraffes within the wild—changing into the primary Western researcher to take action. Her discovery was easy however profound: the sample of spots on a giraffe is as distinctive as a fingerprint, permitting scientists to re-establish particular person animals over time.
At this time, practically seventy years later, that perception is powering one of the superior AI conservation instruments ever created.
We’re proud to announce the launch of GIRAFFE (Generalized Picture-based Re-Identification utilizing AI for Fauna Function Extraction)—an open-source AI software developed by Microsoft’s AI for Good Lab, in partnership with the Wild Nature Institute, to assist conservationists monitor and shield Tanzania’s endangered giraffe inhabitants. Constructed over a decade of collaboration, this software is already serving to to supply vital insights for conservationists to assist stabilize key giraffe populations within the area, and giraffes are only the start.
Why it issues
Giraffes are discovered solely in Africa—and in Tanzania, their numbers have declined by greater than 50% within the final 30 years. Grownup females are sometimes focused by poachers, leaving populations susceptible and fragmented. Conservationists want knowledge to reverse this pattern: survival charges, migration routes, copy patterns, and extra. However gathering and analyzing that knowledge has historically required monumental handbook effort.
That’s the place AI will help.
What GIRAFFE does
Our new software—GIRAFFE—makes use of pc imaginative and prescient to establish particular person giraffes primarily based on their spot patterns, additionally facilitates automated particular person annotations and dataset curation. It’s quick, scalable, and designed for real-world conservation:
- AI-Powered Recognition: Matches giraffes with over 90% accuracy, reaching 99% in lots of instances.
- Finish-to-Finish Workflow: Helps each step of the method, from photograph uploads to skilled evaluate to seamless catalog updates.
- Accessible Design: A clear, user-friendly interface permits each scientists and area researchers to make use of it—no coding required.
- Scalable Infrastructure: Handles hundreds of photos shortly, processing every match in below two seconds.
- Open Supply: Out there now on GitHub, with instruments for each technical and non-technical customers.
Each time a giraffe is photographed—often the proper aspect, which serves as its “ID card”—the software compares the picture to a catalog and both identifies the person or flags it for skilled evaluate. Every survey generates over 1,500 photos; GIRAFFE turns what used to take days into minutes.
“Sample matching software program and pc imaginative and prescient has allowed us now to maintain monitor of hundreds of particular person giraffes. We take photographs of each giraffe we see, and we feed them right into a sample recognition software program, which kinds the premise of all of our knowledge that we use to know the place they’re doing nicely, and if they aren’t doing nicely, why—and we are able to develop efficient conservation actions,” stated Derek Lee and Monica Bond at Wild Nature Institute.”
A mannequin for broader conservation
GIRAFFE’s structure isn’t only for giraffes. It may be tailored for any species with distinctive visible patterns—zebras, tigers, whale sharks, and extra. That’s the facility of open science: we construct as soon as, and the advantages multiply.
We’re particularly grateful to the Wild Nature Institute and the Masai Giraffe Conservation Venture. Their tireless fieldwork, deep experience, and shared dedication have made this progress doable.
At Microsoft, we consider that a few of the world’s largest challenges—local weather, biodiversity, sustainability—require not solely dedication, however collaboration. AI alone gained’t save the giraffes. However within the arms of devoted scientists, it could make a world of distinction.
Let’s be sure that the tallest land mammals on Earth nonetheless have a spot to face.