Curved image sensors and their fabrication technologies for high-performance artificial vision systems

Bioinspired visual devices based on curved image sensors have attracted extensive attention due to their advantages of simplified optical structure, reduced data redundancy, and ultrawide field-of-view imaging capability. Such features endow them with great potential in emerging vision‐dependent applications, including autonomous driving, humanoid robotics, drones, and intelligent navigation systems. However, the fabrication of high-performance curved image sensors remains extremely challenging, primarily because conventional planar micro–nano manufacturing processes are incompatible with large-curvature substrates. The formation of high-density photodetector arrays on curved geometries requires precise control over material deformation, interfacial stress, and electrical connectivity, all of which go beyond the capabilities of existing lithographic techniques. Consequently, the development of reliable and scalable integration strategies for curved pixel arrays has become an urgent research focus in recent years.This review systematically traces the major advances made in curved image sensor technology over the past two decades. Firstly, the evolution of bioinspired architectures, from the hemispherical electronic eye mimicking the human retina, to compound-eye cameras inspired by arthropods, is analyzed to highlight how bionic design principles contribute to overcoming optical aberrations and improving imaging uniformity across wide viewing angles. Furthermore, we summarized representative fabrication approaches such as mechanical deforming and transfer, stretchable circuit integration, origami–kirigami transformation, and monolithic growth on nonplanar surfaces. Each method is discussed from perspectives of device structure, mechanical modeling, optical performance, and process feasibility. The paper also reviews breakthroughs in flexible optoelectronic materials and photosensitive nanostructure systems that enable large-strain accommodation and efficient light detection on curved substrates. Recent progress in perovskite-based photodetectors, organic thin-film electronics, and two-dimensional semiconductors such as MoS2 and graphene has opened new pathways toward highly integrated, lightweight, and conformable imaging modules. These efforts are paving the way for constructing artificial vision systems capable of matching and eventually surpassing the dynamic range, adaptability, and sensitivity of biological eyes.Finally, key challenges and potential directions for future research are discussed, including large-scale curved array integration, high-resolution integration strategies, optical-electronic co-optimization, and system-level intelligence coupling with neuromorphic computing. By summarizing the historical development, existing technical barriers, and emerging solutions, this work aims to provide researchers with an in-depth understanding and valuable insight into the design and fabrication of next-generation curved imaging sensors, thereby promoting their practical deployment in advanced intelligent visual systems.