SLOOP is a field-tested, non-invasive monitoring platform that identifies individual koalas using nose-pattern recognition, with human-in-the-loop relevance feedback for continuously improving accuracy.
Scroll below to know more
The dorsal nose pattern of each koala is unique and remains stable across its lifetime, providing a reliable biometric marker for individual identification. SLOOP performs photo-identification on this pattern: the nose is localised in a submitted image and matched against a reference catalogue of previously identified individuals, returning the closest candidates for verification. The system is currently in field use on Raymond Island, Victoria.
An identification algorithm compares the nose pattern of a newly photographed koala against reference images in the database and returns a ranked list of candidate matches.
The highest ranked candidate matches are reviewed by a human expert to confirm the individual's identity and resolve ambiguous cases.
Interfaces for the koala identification workflow.
Please log in to use these tools.
Upload a koala image together with sighting information to add a new observation to the system.
Open toolCompare a query image against known koalas in the database and review the most likely matches.
Open toolExplore individual koalas, view all associated images, and see where they have been sighted on a map.
Open toolThe people behind the Sloop Koala identification system.
Lead Investigator
Principal Research Scientist at MIT EAPS and Director, Earth Signals and Systems Group (ESSG). Pioneer in computer vision and machine learning, image and pattern retrieval for individual animal identification in wildline monitoring and conservation.
Researcher
Computer-vision researcher at IPN, Mexico. Contributes to feature extraction, scoring methods, and deployment.
Researcher
Researcher at IPN working on scoring algorithms for koala identification.
Researcher
Field researcher with the Koala Island Foundation, Australia. Bridges conservation field work with system requirements and identity validation.
Researcher
Research affiliate at MIT ESSG working on systems operations and the relevance feedback implementation.
Questions about a submission, or want to help improve koala identification? We'd love to hear from you. Email us at
sloopkoala@mit.edu