Raw manual and AI-assisted segmentation data supporting: AI-assisted image analysis of dominant Antarctic pioneer benthic communities on Autonomous Reef Monitoring Structures (ARMS)

Pioneer benthic taxa are indicators of ecosystem responses, being among the first organisms to react to seasonal and environmental changes. Their settlement links meroplanktonic stages with established benthic communities, providing insight into recruitment dynamics and early community development, a process particularly relevant in the Southern Ocean under ongoing climate-driven change. This dataset provides the raw segmentation output from a study applying the AI-based semantic segmentation software TagLab to quantify pioneer benthic taxa colonizing Autonomous Reef Monitoring Structures (ARMS) deployed in the Ross Sea, Antarctica. Three dominant pioneer taxa, Micropora sp., Camptoplites spp., and Beania erecta Waters, 1904, were analyzed using taxon-specific AI models and compared against manual segmentation. Files:- Manual_segmentation_raw_data.csv: raw output from manual segmentation- AI_segmentation_raw_data.csv: raw output from AI-assisted (TagLab) semantic segmentation This dataset supports the manuscript: 'AI-assisted image analysis of dominant Antarctic pioneer benthic communities on Autonomous Reef Monitoring Structures (ARMS): performance, opportunities and limitations' (submitted to Ecological Processes, Special Issue 'AI-Enabled Ecological Intelligence').