The Hong Kong University of Science and Technology (HKUST) has launched a new interdisciplinary research center focused on the ocean, and simultaneously announced its first major collaborative project: an artificial intelligence system designed to monitor coral reef health at a scale and speed far beyond current manual methods.
According to HKUST, the Center for Marine Intelligence (CMI) brings together expertise in artificial intelligence, robotics, and marine science. The center is led by Professor Yeung Sai-Kit, a professor in the Division of Integrative Systems and Design (ISD) and the founding director of CMI. Associate directors include Professor Rob Scharff, assistant professor in ISD, and Professor Hu Wenqi, assistant professor in the Department of Mechanical and Aerospace Engineering. The center aims to advance ocean observation, autonomous exploration, biodiversity research, and ecological conservation.
CMI's inaugural international research project, named "AI4Coral: Automated Coral Reef Monitoring and Analysis Across In-Situ and Ex-Situ Environments," is a collaboration with the King Abdullah University of Science and Technology (KAUST) in Saudi Arabia, specifically with the KAUST Coral Restoration Initiative (KCRI). The project is co-led by Professor Yeung, Professor David Suggett (Director of KCRI), and Dr. Liz Goergen of KCRI.
Coral reefs support fisheries, tourism, and coastal protection, and HKUST noted that they directly or indirectly benefit more than 10% of the world's population. However, climate change and human pressures are threatening these ecosystems, creating a need for more efficient monitoring solutions.
The core technology behind the project is HKUST's Coral AI model. HKUST reported that, compared with traditional point-sampling tools, the model performs full-image dense segmentation automatically and completes image analysis over 1,000 times faster. For a 500 × 500-pixel image, the model classifies all 250,000 pixels, providing 5,000 times as many classified image locations as a 50-point manual annotation.
The teams plan to further develop the model to support KCRI's coral restoration efforts, with the goal of deploying it directly on underwater edge devices. According to HKUST, the AI model is designed to autonomously identify targeted coral regions in real time and support the calculation of key health metrics, including coral cover, growth rate, species diversity, and bleaching rate. The technology is also set to be integrated with underwater robotic platforms from the KAUST research team.
A preliminary version of the model has already been tested on imagery from the Red Sea and KAUST's in-water and land-based coral nurseries, as well as reef imagery from Hong Kong's Hoi Ha Wan, provided through a collaboration with WWF-Hong Kong. HKUST stated that broader field validation is ongoing.
Professor Yeung said in the announcement: "Coral reef monitoring has traditionally relied on time-consuming surveys and manual image analysis. By combining novel technologies and knowledge across AI, robotics, and marine science, this collaboration aims to enable faster, more scalable, and data-driven assessments of reef health. We hope these technologies will support scientists and conservation practitioners in making more informed decisions to jointly protect and restore coral reef ecosystems."
Professor Suggett added: "We are excited about this collaboration between two world-leading universities, bringing together HKUST's innovations in coral reef analytical platforms and KAUST's expertise in scaling coral conservation and reef restoration. Through this collaboration, HKUST researchers will work alongside KAUST faculty and the KCRI team to advance automated reef health assessment using AI and machine learning, building on existing research strengths and opening further opportunities for joint research."
Beyond the AI4Coral project, CMI is extending its AI capabilities to reconstruct 3D reef structures from field transect videos, with support from HKUST's Ocean Central Research Facility and Sustainable Smart Campus Office. Related coral-monitoring research is also supported by the Environment and Conservation Fund. The team has presented its research at the 16th International Coral Reef Symposium in Auckland and is pursuing further collaboration through a multi-region CORDAP proposal and ongoing discussions with OceanX.
CMI's broader research portfolio includes the Sustainable Smart Marine Grid, the Sustainable Smart Campus Pier Living Lab, multi-domain aerial-surface-underwater robotic platforms, and bio-inspired soft robotic fish and jellyfish drones led by Professor Scharff and Professor Hu. Leveraging HKUST's accessible shoreline, the center integrates AI, robotics, and marine ecology to advance ocean monitoring, STEAM education, and community conservation.
For university students, particularly those studying computer science, marine biology, environmental science, or robotics, this initiative represents a concrete example of how AI and robotics are being applied to real-world environmental challenges. The collaboration between HKUST and KAUST also highlights growing opportunities for international research partnerships in sustainability-focused technology. Students interested in interdisciplinary work combining AI with marine science may find emerging research pathways and potential future projects through CMI and its partners.