Help the Bureau of Ocean Energy Management (BOEM), NOAA Fisheries, and Wild Me accurately identify endangered Cook Inlet beluga whales from photographic imagery. Scalable photo-identification of individuals is critical to population assessment, management, and protection for these endangered whales. #climate
Beluga whales are extremely sociable mammals that live in pods and are known for being one of the most vocal of all whales. Measuring up to 15 feet and 3,500 pounds, they are found in seasonally ice-covered waters throughout the arctic and sub-arctic regions of the northern hemisphere. Belugas are legally protected in the United States under the Marine Mammal Protection Act (MMPA). The federal agency National Oceanic and Atmospheric Administration (NOAA) Fisheries monitors five different populations of belugas across Alaskan waters, with a focus on the Cook Inlet belugas. This depleted and endangered population began declining in the 1990s from overhunting and currently contains less than 300 surviving individuals.
Cook Inlet beluga whales were listed as an endangered species under the Endangered Species Act in 2008 and are at risk for extinction. The Marine Mammal Laboratory at the NOAA Alaska Fishery Science Center began conducting an annual photo-identification survey of Cook Inlet belugas to more closely monitor and track individual whales. Specifically, the Lab takes overhead photographs of these belugas using drones and lateral photographs using vessels to estimate abundance in the population.
Today, processing and analyzing new survey images of Cook Inlet belugas is largely manual and consumes significant time and resources. Photographs are visually matched to existing images based on whale features like color, unique marks and scratches, and dorsal ridge shape. For instance, the position or number of marks on a whale may provide clues about its identity. New and improved methods are needed to help automate this process and accurately find matches of the same individual whale in different survey images.
The goal of this challenge is to help wildlife researchers accurately identify endangered Cook Inlet beluga whale individuals from photographic images. Specifically, the task is query ranking for images of individual whales given a query image against an image database, which is a key step in the full photo-identification process.
The Bureau of Ocean Energy Management (BOEM) is looking for an automated solution to integrate into the existing Flukebook platform. Accelerated and scalable photo-identification of individuals is critical to effective population assessment, management, and protection for the Cook Inlet belugas. Machine learning has the potential to expedite the creation and analysis of large datasets of endangered beluga whale images.
Prizes will be awarded based on the accuracy of model matches, with scores on a private leaderboard. All submissions are scored on the private leaderboard, and each participant's best score on the private leaderboard is used for prize determination. Private leaderboard rankings are not revealed until the end of the competition. Please note that, as always, it is recommended to evaluate your models using best practices like holdout and cross-validation to ensure you are not overfitting to the training data or public leaderboard.
This competition also includes a bonus prize for intermediary explainability outputs which is further described below.
Competition End Date:
June 30, 2022, 11:59 p.m. UTC
Note on prize eligibility: The term Competition Sponsor in the Competition Rules includes the Bureau of Ocean Energy Management as well as all federal employees acting within the scope of their employment and federally-funded researchers acting within the scope of their funding. These parties are not eligible to win a prize in this challenge.
Bonus Awards: Algorithm Explainability
The top four winners ranked by performance on the private leaderboard will also be invited to participate in the bonus round after the close of the competition. To submit to the bonus round, the invited winners will modify their winning code submissions to additionally output explanations for their models' predictions, such as bounding boxes that highlight associated regions-of-interest (ROIs) between the query image and database images. A judging panel of subject matter experts will review the bonus submissions and determine whether each meets the requirements for the bonus prize. All bonus round participants who successfully meet the requirements will evenly split the bonus round prize pool.
How to compete
- Click the “Compete” button on the sidebar to enroll in the competition.
- Get familiar with the problem through the overview and problem description. You might also want to reference additional resources available on the about page and the user forum.
- Download the data from the data tab.
- Create and train your own model. The benchmark blog post is a good place to start.
- Package your model files with the code to make predictions based on the runtime repository specification on the code submission format page.
- Click “Submissions” on the sidebar followed by “Make new submission” to submit your code as a zip archive for containerized execution. You’re in!
- We will generate outputs for the evaluation dataset and score your submission in our cloud environment.
This challenge is sponsored by our friends at the Bureau of Ocean Energy Management.
Banner image courtesy of Paul Wade/NOAA Fisheries.