Microsoft, in collaboration with the Wild Nature Institute and the AI for Good Lab, unveiled results from their GIRAFFE project. This project is designed to rescue Tanzania dwindling giraffe populations through state‑of‑the‑art artificial intelligence.
Africa’s giraffe numbers have plummeted due to habitat loss and illegal hunting. With over 3,000 giraffes tracked in northern Tanzania, gaining precise insights into births, deaths, and movements has become urgent
GIRAFFE: AI-Driven Individual Tracking
GIRAFFE [Giraffe Identification & Recognition Using AI Facial Encoding] merges computer vision and pattern‑recognition to monitor each giraffe’s unique coat pattern across camera trap and drone images. Using Azure tools, the model detects giraffe torsos and key points such as head, neck, and legs with high accuracy and speed
Previously, identifying individuals manually could take days or weeks. GIRAFFE now automates cropping, recognition, and measurement, delivering results in minutes by leveraging Azure’s machine‑learning pipelines .
Field-Tested and Scalable
Built in Microsoft’s AI for Good Lab in Kenya, and co‑developed with Wild Nature Institute. GIRAFFE processed tens of thousands of images spanning thousands of giraffes. Early field deployments show near‑perfect pattern recognition, even when torsos are partially obscured.
Echoing Microsoft’s broader biodiversity efforts such as SPARROW, GIRAFFE is being released as open‑source. Conservationists worldwide can now leverage this technology to adapt the tool to their local populations and species
Transforming Conservation Strategies
By integrating GIRAFFE with satellite and acoustic data. Conservationists gain near real‑time insights into giraffe population health, enabling rapid interventions against threats like poaching or habitat degradation. This data empowers informed policy and community action.
Over the next year, GIRAFFE will scale across Tanzania, with plans to adapt it for other species with unique patterns [e.g. elephants, zebras, whales]. Integration into broader AI‑powered conservation platforms could help create a holistic, data‑driven approach to saving wildlife globally.
