Metasurface-enabled imaging: From optical design to intelligent information processing

Metasurfaces have emerged as a transformative optical platform capable of manipulating amplitude, phase, and polarization at subwavelength scales, enabling unprecedented control of light within ultrathin planar architectures. Although early developments largely focused on replacing conventional bulk optical components with planar analogues, recent advances have expanded the role of metasurfaces beyond wavefront engineering to functional imaging systems integrating optical encoding, computational reconstruction, and artificial intelligence. This review provides a systematic overview of metasurface-enabled imaging as an interdisciplinary framework combining nanophotonic device engineering, computational optics, and intelligent information processing. Within this framework, metasurfaces serve as enabling optical components for wavefront manipulation, image formation, and information encoding, while operating alongside illumination, detection, electronic, and computational subsystems. We critically examine the physical principles governing metasurface operation, advanced simulation and inverse-design methodologies, scalable nanofabrication strategies, and imaging architectures spanning metalens-based, coded-aperture, lensless, and hybrid computational systems. Emphasis is placed on optical encoding mechanisms, compressive sensing, deep-learning-based reconstruction, and end-to-end hardware–algorithm co-design for intelligent and task-specific imaging. By connecting advances in photonics, computation, and machine intelligence, this review establishes a system-level perspective on the development of compact, adaptive, and highly integrated imaging technologies.