Leveraging AI to enhance the lives of Toronto Zoo residents

The Challenge

Current methods in animal research can be lengthy and tedious and typically involve the collection of real-time data via paper, pen, and clipboard, or videos that must be manually annotated

The Solution

Newly adapted for use at the Toronto Zoo with a focus on Sumatran orangutans, a first-of-its-kind AI model is unobtrusively and non-invasively capable of mass data collection, individual identification, pose estimation, 24/7 real-time behaviour monitoring, benefiting both animals and zookeepers alike

By harnessing the power of AI, six orangutans at Toronto Zoo collaborate with researchers from York University to enrich habitats and quality of life

As more and more species are placed at serious risk of extinction, wildlife conservation’s critical importance is growing. The study of animal behaviour is essential in helping us create optimal species-specific environments where animals, especially those that are endangered, can live happy and healthy lives while also shedding a light on human behaviour. As zoos and other conservation spaces work to create habitats that enrich animal welfare, monitoring the health and well-being of their animal residents is imperative in being able to achieve this.

However, actual methods for collecting this type of data can be tedious and lengthy. Research on animal behaviour is typically conducted manually and in real time via paper, pen, and clipboard, or through videos that must later be individually annotated. Despite many technological advancements in the field, current devices capable of recording location, heart rate, and movement must often be physically attached to the animal being monitored. Until now, there has not been a technology or device capable of doing all of this unobtrusively and non-invasively.

In collaboration with Canadian tech company EAIGLE and the conservation-focused Toronto Zoo, researchers from York University have created a technology capable of mass data collection and 24/7 monitoring of the animals in their care. Originally created to help retail stores and attractions adjust their layouts according to customer behaviour as well as for temperature monitoring for COVID-19, the AI model has been adapted by the team to improve animal welfare and quality of life, beginning with six Toronto Zoo residents.

Humans of the forest  

Ranging in age from 16 to 54, Sumatran orangutans Puppe, Ramai, Sekali, Budi, Kembali, and Jingga are part of the first cohort of mammalians benefiting from the first-of-its-kind technology. Expected to eventually move to a new outdoor habitat, the welfare of these orangutans is being monitored to make the transition as smooth as possible and to ensure that they adapt well to their new surroundings.

Phases one and two of the project consisted of training the AI model to be able to detect orangutans in the first place, and then how to recognize them. Now, it is in its third phase.

“After we installed cameras into their habitat, we spent hundreds of hours going through image after image and annotating them,” explains Dr. Jenna Congdon, a postdoctoral fellow on the project. “This was tedious but essential in allowing us to train the AI model to recognize each orangutan individually, but it was also a necessary step before being able to move into phase 3 and collect data on their actual behaviours.”

Given the close physical similarity between orangutans and humans as great ape species, it was a natural decision to first try and adapt this cutting-edge technology to work on orangutans or, as they are known etymologically, the humans of the forest.

“The technology was already there. We just needed to find a way to adapt it to other species. And since orangutans are more like humans and I’d been working with them already for many years, it’s what made the most sense,” says Dr. Suzanne MacDonald, a professor in the Department of Psychology at York University and the academic supervisor on the project.

The “gold standard”

With the third phase of the project underway, the team has already completed the pose-estimation process (for example, specifying body landmarks like the shoulders and head) and is about to begin behaviour monitoring of the 6 orangutans.

“This is the most exciting part,” says Dr. Congdon. “Being able to analyze what the orangutans are actually doing, how they are using their space, will be incredibly important in further improving their welfare and quality of life.”

“With this data, the zoo will potentially be able to design better habitats in the future, provide enhanced and enriching environments for its residents, and detect disease at an earlier stage,” explains Doctor of Veterinary Medicine, Mina Hosseini, the second postdoctoral fellow on the project.

As the team continues to develop and tweak this technology, they hope it will become the “gold standard” for conservation-focused zoos worldwide.

“Now that we have done the work to adapt this technology to benefit the orangutans, we will be able to accelerate the process and expand it to other species,” says Mahdi Marsousi, Chief Technology Officer at EAIGLE. “It has the potential to benefit not only the Toronto Zoo in its conservation efforts, but also zoos and conservation spaces worldwide.”

“In the end, it’s all about enriching the lives of the animals,” adds Maria Franke, Manager of Welfare Science at the Toronto Zoo.

Canada at the forefront 

While the main purpose of the project is to ultimately program artificial intelligence to recognize specific behaviours, it will also allow for zoos, conservation sites, and researchers to collect data unobtrusively in an efficient and unbiased manner.

“Canada is at the forefront of this type of research,” explains Dr. MacDonald.

“There is a technology that’s capable of individual recognition and thermal recognition. But up until now there hasn’t really been something that does all this combined with behaviour to the specificity that we’re trying to achieve here. And it’s happening right here in our backyard. And what better place to do it than the Toronto Zoo?” adds Dr. Congdon.

As the team continues to work on testing and refining their technology, they credit Mitacs with providing the guidance and support they needed at a critical juncture of development.

“Important projects like this deserve the spotlight so that we can remind people of all of the good that happens at the Toronto Zoo and the constant care that we’re providing for these animals,” says Dr. Congdon. “These are their homes and we’re trying to make those homes as comfortable as possible, and support from Mitacs has been critical in helping us achieve that.”

“From the Toronto Zoo side, it has been completely seamless, and we want to continue this amazing collaboration,” adds Franke.

“The only reason we have these two fabulous postdoctoral fellows is because of the essential support and funding we’ve received. We would simply not be doing this if it wasn’t for Mitacs,” concludes Dr. MacDonald.

Mitacs’s programs receive funding from valued partners across Canada. We thank the Government of Canada, the Government of Alberta, the Government of British Columbia, Research Manitoba, the Government of New Brunswick, the Government of Newfoundland and Labrador, the Government of Nova Scotia, the Government of Ontario, Innovation PEI, the Government of Quebec, the Government of Saskatchewan, and the Government of Yukon for supporting us to foster innovation and economic growth throughout the country. 

Do you have a business challenge that could benefit from a research solution? If so, contact Mitacs today to discuss partnership opportunities: BD@mitacs.ca.

Mitacs Team

Mitacs’s website content is created by people throughout our organization, united in their passion for innovation and eager to share their perspectives with others in the innovation ecosystem. 

More posts by this author Contact Me