AWS Partner Capgemini Builds an AI Tool to Identify Sperm Whales with 97.5% Accuracy

AWS Partner Capgemini builds innovative artificial intelligence and machine learning solutions for its customers by being a member of the AWS Partner Network. The technology services company became an AWS Partner in 2012 and is a member of multiple AWS Partner programs, including the AWS Well-Architected Partner Program. Through its annual Global Data Science Challenge (GDSC), Capgemini empowers its team to develop artificial intelligence and machine learning solutions to address single use cases. During the 2020 GDSC, the winning team used AWS services to build a sperm whale identification tool that matches sperm whale flukes with 97.5 percent accuracy in minutes. 

For technology services company Capgemini, developing solutions to the world’s most pressing environmental problems is a top priority. It builds those innovative cloud-native solutions using Amazon Web Services (AWS) and is an AWS Partner. For its 2020 Global Data Science Challenge (GDSC), an event during which internal teams build solutions for real-life use cases, Capgemini worked with Whale Watch Azores marine biologist Lisa Steiner. Teams competed to develop an artificial intelligence (AI) model that would match and analyze sperm whale flukes. Using AWS Partner Network resources and AWS services such as Amazon SageMaker, a machine learning service used to build, train, and deploy machine learning models for virtually any use case, the winning team developed a scalable, automated solution to more rapidly and accurately identify and help protect the vulnerable species.

Founded in 1967, Capgemini provides technology, consulting, and digital transformation services across industries, including the public sector, manufacturing, and financial services. It oversees a team of 289,000 people in over 40 countries. Sustainability is a driving force behind many of its solutions. “Capgemini is committed to environmental sustainability and reversing climate change,” says Kristin O’Herlihy, Global Marketing Lead for the Insights & Data Business Unit at Capgemini. 

Capgemini began to use cloud technologies in 2008. “We evaluated the best workloads in the early days of the cloud, then gradually worked toward more complex, critical workloads,” says Nico Steenkamp, Global Amazon Partner Executive at Capgemini. “As an organization with a diverse portfolio, Capgemini has always recognized the value of partnership programs.” The company became an AWS Partner when the AWS Partner Network launched in 2012. It is a member of seven AWS Partner programs, including the AWS Solution Provider Program, which enables the company to manage, service, support, and bill AWS accounts for its customers. As a member of the AWS Well-Architected Partner Program, Capgemini can access the resources it needs to help AWS customers establish good architectural habits, reduce risk, and respond faster to changes that affect designs, applications, and workloads. 

Capgemini was inspired to become an AWS Partner to access funding opportunities, training programs, and the ability to gain AWS Competencies, which help differentiate the company from its competitors. As of June 2021, Capgemini has over 10,000 employees trained on AWS and over 3,500 employees who have AWS Certifications, including employees designated as AWS Certified Solutions Architect – Professional, an AWS Certification for individuals who perform a solutions architect role and have two or more years of hands-on experience managing and operating systems on AWS. 

Being an AWS Partner has provided Capgemini the resources it needs to deliver innovative, scalable, and flexible solutions. “We need global access to consistent content, people, and services,” says Steenkamp. “The AWS Partner Network is beneficial because we have the full spectrum of programs available to us on demand and globally.”

Because Capgemini is a member of the AWS Public Sector Partner Program, which recognizes AWS Partners that have cloud-based solutions and experience across government, space, education, and nonprofits, the GDSC often focuses on issues related to public sector research. The company chose whale conservation as the theme for its 2020 competition. While searching for a relevant use case, the Capgemini team was introduced to Steiner, a marine biologist who has conducted sperm whale research for over 30 years. Steiner, who operates out of Portugal, has thousands of sperm whale images. Sperm whales can be identified by the shape of their flukes, or tails, which are as unique as fingerprints. 

However, matching flukes was a time-consuming process. “Lisa worked with an older software program that could do some matching but required intensive manual preprocessing. It took ages to match a new image to older pictures,” says Daniel Kühlwein, Data Scientist at Capgemini. “And there was no clear way to compare images with other whale researchers.” The Capgemini team realized that by using Steiner’s photos, it could create an integrated solution to help researchers better understand migration patterns, genealogical lines, and other information that provides insight into the sperm whale’s way of life. Marine biologists could use this data to inform environmental policies and strengthen protections for the species, which the International Union for Conservation of Nature lists as endangered.

During the 2020 GDSC, Capgemini employees built AI-based solutions to help Steiner streamline the fluke matching process. The winning team used AWS services such as Amazon SageMaker and Amazon Simple Storage Service (Amazon S3)—an object storage service that offers industry leading scalability, data availability, security, and performance—to develop a deep neural network that can identify flukes with 97.5 percent accuracy. With this solution, Steiner can upload new pictures, and the automated processing pipeline will match, analyze, and report information on individual sperm whales within a few minutes.

Using AWS services, Capgemini deploys deep learning solutions at scale across thousands of users. Its solutions have the potential to innovate several environmental research processes, including whale identification and ocean monitoring. “We can deploy highly specialized Capgemini employees who have extensive training on using the latest AWS services,” says Steenkamp. “We also bring our deep industry knowledge and engineering capabilities to the table. Combined with our cloud expertise, our position as an AWS Partner differentiates us from our competitors.” Since joining the AWS Partner Network, Capgemini has seen a 62 percent year-over-year increase in bookings for its AWS Partner services. 

As an AWS Partner, Capgemini has the resources and support it needs to deliver innovative solutions to its customers. The company is expanding its sperm whale solution to include other species, such as humpback whales. It also hopes to enable greater collaboration by creating a centralized, scalable database where whale researchers can upload images and better identify individual animals. To generate awareness around greener information technology, the company is launching a co-innovation program called Responsible Earth with the support of AWS. “The AWS Partner Network team has been very involved in helping us drive Responsible Earth into the marketplace,” says Steenkamp. “Through this program, we aim to develop solutions that not only are useful for our customers in traditional ways but also drive sustainability and new lines of business.