Abstract： Historical and modern astronomical observations have accumulated a vast amount of image data. In recent years, with the rapid development of AI, the application of computer vision (CV) techniques to astronomical images has accelerated significantly. Along with the next generation of astronomical surveys, this growing demand may lead to an “AlphaFold moment” for astronomical imaging. In this talk, I will first highlight the key differences between astronomical images and traditional CV data, including low signal-to-noise ratios, measurement uncertainties, and masking effects. These characteristics pose unique challenges for CV applications. Motivated by scientific needs, we have carried out a series of studies leveraging AI for efficient computation and feature extraction, including the recognition of targets in historical photographic plates, and the use of deep features and super-resolution models to investigate galaxy structures. These examples demonstrate the strong potential of AI in advancing astronomical research. Finally, I will discuss future directions and emerging demands for AI in astronomical image analysis.
Bio： 沈世银，中国科学院上海天文台研究员、博士生导师。自1998年毕业于重庆大学后，先后在上海天文台攻读硕士（1998–2001），并在德国马普天体物理研究所完成联合培养博士（2001–2004）。自2004年起在上海天文台工作至今。主要从事星系天文学研究，现任载人空间站巡天望远镜（CSST）科学数据系统副主任设计师，负责CSST-IFS数据处理软件开发工作。曾两次获得上海市自然科学二等奖，入选上海市优秀学术带头人，并入选2021–2025年爱思唯尔中国高被引学者。
