Researchers automate detection of fish cleaning interactions

According to the Nature preprint, the authors developed a semi-automated system to detect cleaning interactions between the cleaner wrasse (Labroides dimidiatus) and the powder blue tang (Acanthurus leucosternon) in a controlled three-dimensional laboratory setting. The pipeline used **DeepLabCut** for markerless pose estimation and a downstream classifier built on the tracking data. Per the preprint, the classifier detected cleaning interactions with **90% accuracy**, misclassified about **15%** of non-interactions as interactions, and flagged **25%** of footage as containing interactions, thereby reducing manual annotation by **75%**. The manuscript is an unedited preprint version published on Nature on 03 June 2026. According to the Nature preprint published 03 June 2026, the authors report a semi-automated pipeline that tracks and classifies cleaning interactions between the cleaner wrasse (Labroides dimidiatus) and the powder blue tang (Acanthurus leucosternon... [1951 chars]