Since 2019, they have operated a facial recognition tool prototype that helps identify individual orangutans in their habitat in the Punggualas area of Sebangau National Park, Central Kalimantan, in less than 10 minutes. In the past, the process usually took days.

“We’ve applied this technology to orangutans because we have a large database of them,” WWF Indonesia finance and technology director Aria Nagasastra told an online press conference on Thursday.

The images of orangutans are later gathered using mobile phones and motion-activated cameras in their basecamp and uploaded to Amazon Simple Storage Service (Amazon S3) for analysis. The process, which involves the use of cloud machine-learning platform Amazon SageMaker, allows data scientists and developers to build, train and deploy machine-learning models. Among the benefits are the ability to increase data accuracy by up to 95 percent to determine whether the orangutans are pregnant, ill or injured.

“If we can identify orangutans, we’ll be able to identify their communities and observe their behaviors as individuals and part of a group,” Aria said. “Later on, we’ll be able to manage their habitat and provide the best place for them.”

According to Thomas Barano, head of conservation science at WWF Indonesia, the team had so far identified five orangutans. 

Identifying these primates is quite challenging given that inserting microchips under the skin is not an option, as the orangutans are known for attempting to remove them, which risks injury. Tracking them through fingerprints is also difficult due to the density of the jungle and their tendency to hang on trees.

Additionally, capturing quality images consistently poses challenges in terms of finding the proper angle and sufficient sunlight.

Read also: More than meets the eye in photo of orangutan ‘offering help’ to man

Regarding data security, AWS utilizes shared responsibility, which means the company is responsible for the infrastructure that operates all services in the AWS Cloud, while the user’s responsibility depends on the services they use.

Aria said the technology could also benefit other orangutans’ observation base camps, considering the team's success in using the prototype and gaining support from donors and local administrations.

“We’re optimistic that various conservation activities can be done using this machine-learning tool; [enabling us] to detect and identify wild animals, monitor the possibility of wildlife transaction, observe wildlife transactions on social media and measure plastic pollution distribution in the ocean,” Aria said.

Plans are in store to use additional machine-learning services, such as Amazon Rekognition, to improve the speed and accuracy of population identification and tracking efforts.

Orangutans are known for their intelligence as they are able to arrange objects and live in groups in which each orangutan exhibits different characteristics. However, their population is currently threatened by illegal hunting, habitat destruction and wildlife trade.

According to WWF data, the orangutan population in Kalimantan has declined more than 50 percent in the last 60 years.

Since 2005, WWF Indonesia has conducted assessments on the orangutans’ well-being and built a conservation area at Sebangau National Park.

Established in 2006, AWS boasts more than 175 services for areas such as computing, storage, database, the Internet of Things (IoT), machine learning and artificial intelligence. (kes)

Popular

To my horror, it’s just a movie!

RI, Australia strengthen partnership on taxing crypto assets

Pentagon chief congratulates Prabowo on being elected President