Despite the rapid increase in imagery collected in marine environments the time it takes to manually process it for biological, geological, and ecological information remains a major bottleneck in advancing our understanding of these environments. Advances in artificial intelligence (AI), particularly in computer vision and machine learning, have the potential to change how we process imagery and increase our ability to make insights and discoveries via marine imagery. We invite contributions that demonstrate the use of AI techniques—such as object detection, image segmentation, unsupervised classification, and self-supervised learning—for automating the analysis of seafloor imagery, video, and sensor-integrated data. Topics may include, but are not limited to, automated species detection and counting, habitat classification, anomaly detection, mapping and photogrammetry of marine environments, training datasets and annotation strategies, and validation methods. We also welcome presentations that address the limitations, ethical implications, and deployment of AI systems in remote marine settings.

This session aims to foster interdisciplinary collaboration and discuss future directions for AI-enabled exploration and monitoring of marine environments and help identify factors holding back its implementation at large scales. Scientists, data scientists, engineers, and marine managers are encouraged to contribute, those advancing open data, community-driven model development, and field-based validation efforts.

• Deep Sea Exploration -
• Ocean Technologies and Observatories -