Data Annotator
I was involved in annotating bounding boxes on salmon fish images to detect instances of fish damage. The main objective was to develop AI models capable of identifying and analyzing various types of damage, such as Snout, Wound, Operculum, Exophthalmos, Eye clouding and bleeding and fin damage, to improve fish health monitoring in aquaculture facilities. Over the course of the project, I annotated a dataset consisting of thousands of salmon fish images, encompassing a diverse range of damage scenarios and image conditions. The labeling tasks involved not only identifying the damaged areas but also accurately delineating their boundaries with bounding boxes to facilitate subsequent analysis. In terms of project size, the dataset was substantial, requiring several months of dedicated effort to complete the labeling tasks.