Low-Dimensional Perovskites for Neuromorphic Vision Computing

Neuromorphic computing offers a promising route to overcome the energy and data-movement limitations of von Neumann architectures, particularly for in-sensor and edge-intelligence applications. Achieving such systems relies on functional materials that intrinsically integrate sensing, memory, and computation. Metal halide perovskites have emerged as a compelling platform due to their outstanding optoelectronic properties, tunable low-dimensional structures, and pronounced ion-migration dynamics. This review focuses on low-dimensional perovskite nanostructures for neuromorphic vision devices. We summarize recent advances in the synthesis and integration of zero-, one-, and two-dimensional perovskites, and analyze ion-migration-driven optoelectronic and memristive mechanisms underlying synaptic and neuronal behaviors. By correlating material and device characteristics with neural network algorithms, we discuss pathways toward efficient neuromorphic computing. Finally, representative opportunities for perovskite-based in-memory, in-sensor, and near-sensor computing are discussed, highlighting key challenges toward monolithic sensing-memory-computing integration.