The technology behind the African Carnivore Wildbook makes the process of identifying individuals easier, faster, and more accurate, not only freeing up valuable time and resources for conservationists but allowing citizen scientists from all over the world to become part of the equation by quickly and easily providing images and other information to the African Carnivore database. The machine learning algorithms that power ACW deliver two important functions that make this possible.
Anyone who’s been to Africa knows how well these animals can blend into their environment, making them very difficult to spot. ACW’s technology can find an animal or animals within an image, then isolate them by putting a digital ‘box’ around each individual. Once the system has detected an animal, it can begin the process of identifying it.
Prior to the emergence of the technologies delivered by ACW, conservationists had to spend countless hours laboriously comparing images manually to try to identify individual animals. The Wildbook's machine learning platform automates and speeds that process by drilling into a massive database and attempting to match the unique physical characteristics of an individual animal with those of animals previously identified. When the ACW comparison results in a match, the system offers up the results for analysis automatically. If no match is made, researchers know that a new individual has been spotted, and that animal is named and added to the system.