Any-fin is possible with citizen science

CLOSED: This discussion has concluded.

Bruvnet SoftwareWritten by Compliance division

As part of Innovation Month 2020, the Supervising Scientist Branch ran an online fishing competition. The goal was to train their new artificial intelligence model, BRUVNet.

The innovative model helps to automate the identification of fish species in Kakadu National Park billabongs. This involves counting and labelling fish from video footage. The challenge is that BRUVNet needs a significant dataset of labelled fish imagery. A large dataset will allow enough accuracy for regular monitoring.

The challenge saw many individuals contribute to labelling the fish. It was Brooke Batham, Brooke from the department’s Innovation Hub who succeeded in labelling the most. A staggering 943 fish!

‘I think it's so important to get involved in citizen science projects. I always love using the skills I learnt from my marine biology background, and I'm so glad I could help out! Brooke said.

‘This gives me the opportunity to use the skills I have learnt to help protect Kakadu without ever leaving Canberra.’

Even though the challenge has ended, you can still help by taking part online. This is an opportunity to contribute to the protection of Kakadu National Park.


Share on Facebook Share on Twitter Share on Linkedin Email this link
<span class="translation_missing" title="translation missing: en.projects.blog_posts.show.load_comment_text">Load Comment Text</span>