A team of US researchers has developed a new deep-learning-based technique that can automatically identify, count and describe animals in their natural habitats without human intervention. The new technique, reported in the Proceedings of the National Academy of Sciences (PNAS) journal, relies on motion-sensor cameras that photograph the natural ecosystem. It managed to identify animals in 99.3 percent of the photographs at the same 96.6 percent accuracy rate of crowd-sourced teams of human volunteers. The Phys.org website cited Jeff Clune, the senior author of the paper from the University of Wyoming and Ubers Artificial Intelligence Labs, who said: “This technology lets us accurately, unobtrusively and inexpensively collect wildlife data, which could help catalyze the transformation of many fields of wildlife biology.” “This will dramatically improve our ability to both study and conserve wildlife and precious ecosystems,” the German news agency reported. This study obtained the necessary data from Snapshot Serengeti, which has deployed a large number of "camera traps" in Tanzania that collect millions of images of animals in their natural habitat. The new technique has so far harnessed 3.2 million labeled images featuring wild animals such as lions, tigers and elephants. Margaret Kosmala, who works with the Snapshot Serengeti projects, said: “Not only does the artificial intelligence system tell you which of 48 different species of animal is present, but it also tells you how many there are and what they are doing. It will tell you if they are eating, sleeping or moving.” “We estimate that the deep learning technology pipeline we describe would save more than eight years of human labeling effort for each additional 3 million images. That is a lot of valuable volunteer time that can be redeployed to help other projects,” Kosmala added.