A non-profit based in Portland developed an AI computer vision system intended to recognize animals and help conservation groups protect them. Animals like zebras, giraffes, lions, and whales are often endangered, and both conservation groups and scientists need to track these animals. As reported by The Seattle Times, the nonprofit Wild Me’s AI is capable of recognizing animals by their patterns, like spots or stripes, enabling researchers to track animals more effectively.
The process of gathering identifiable information about animals has long been expensive, laborious, and invasive. However, a computer vision application that can recognize animals by sight can make recognizing individual animals much easier. Animals come in all different shapes, sizes, and patterns. If there’s one thing neural networks excel at, it’s recognizing patterns.
Thanks to advances in computer vision and photography, photo surveys are becoming the method of choice for animal population estimates and some forms of animal tracking. The process is far less invasive and much cheaper than traditional methods, not to mention much quicker. Wild Me’s AI is capable of recognizing animals based on unique patterns like a Giraffe’s spots, and it can recognize these specific animals much quicker than a human observer could.
The AI model developed by Wild Me is trained on large volumes of image data. This image data comes from regular citizens as often as it does form scientists. The AI model then uses computer vision techniques and machine learning systems to recognize specific giraffes or other animals.
The new AI system has been well received by conservation scientists. Christin Khan, an aerial surveyor of whales for National Oceanic and Atmospheric Administration (NOAA), was quoted as saying that an AI solution for monitoring endangered species of whales has been wanted for years. Khan works for NOAA tracking North Atlantic right whales, and it’s estimated that there are only about 400 of these whales left. Tracking them proves difficult because they tend to migrate long distances. Wild Me assembles catalogs of images of a given species into a database called a Wildbook. The Wildbook for whales encourages researchers around the globe to collaborate and share their findings and observations, which can make monitoring the movement of whales over large distances easier.
According to Michael Brown, a conservation scientist at the Smithsonian Conservation Bio Institute and the Giraffe Conservation Foundation, the Wildbooks and AI algorithms developed by Wild Me help protect animals in various ways. Said Brown to the Seattle Times:
“We can use this information to track diseases and poaching threats, look at manifestations of diseases. It lets us piece together an understanding of how these threats to giraffes are spatially situated (and) how the giraffes are utilizing different landscapes over time.”
Wild Me has also designed an AI system that pulls data from YouTube videos. The AI analyzes video of marine animals like sea turtles and whale sharks, and it uses this data to get a better estimate of the animal population. This is a useful way of harvesting more data and supporting the creation and optimization of Wildbooks.
The NOAA is employing AI to protect animals in another fashion as well. NOAA scientists are collaborating with Microsoft to design an AI that can monitor aquatic and arctic animals like beluga whales, polar bears, and ice seals. The AI tools are will be trained on sound and be able to distinguish the sound of a seal from the noise a dredging machine makes. An AI will also be trained on images in order to allow airplanes to fly over stretches of sea ice and take counts of polar bears and seals.