Using AI to protect one of the most elusive animals on Earth

Snow Leopards are a highly threatened species, native to the steppes and mountainous terrain of Asia. There are only an estimated 3,900 to 6,500 individuals left in the wild. Their population is a key indicator for the health of the whole ecosystem, but they can be difficult to spot and track in the wild.

Due the cats’ remote habitat, expansive range and extremely elusive nature, researchers use motion-triggered camera traps to observe snow leopards in the wild. Since the cameras trigger on any type of movement, most of the images are of goats, birds, and grass blowing in the wind. Only about 5 percent of the pictures contain a leopard, which can be hard to spot due to their camouflage.

Over the years, more than 1 million images have been captured, but the classification process has been time-consuming. Manually reviewing all images to find a snow-leopards could take thousands of hours of time.

To solve this problem, the Snow Leopard Trust worked with Microsoft to build an image classification model using Azure Machine Learning, which includes Microsoft Machine Learning for Apache Spark. The team additionally created a live dashboard that highlights snow leopard hot spots, or social meeting points for leopards. Rather than taking several hours, automated image classification can now be completed in three minutes, saving researchers over 300 hours per camera survey.