A Flexible Neuromorphic Window Capable of In‐Sensor Computing and Self‐Adaptive Color Regulation

Abstract Smart windows have advanced significantly to mitigate the threats caused by light pollution and promote a comfortable life. However, existing smart windows face challenges in achieving lightweight, high flexibility, portability, and efficient environmental adaptation. Herein, a flexible neuromorphic window based on a feedback loop inspired by the human visual system is presented, capable of in‐sensor computing and self‐adaptive color regulation. The flexible peripheral circuit and the entire neuromorphic window with a size of 10 × 10 cm have masses of ≈1.71 and 9.86 g, respectively. In this loop, optoelectronic synaptic transistors (OESTs) are used as photoreceptors, with a bulk heterojunction conductive channel composed of randomly‐distributed crystallized p‐type and n‐type polymeric nanofibers to promote exciton dissociation. The OESTs exhibited broadband spectral response (≈300‐900 nm), high photosensitivity (≈200 nW cm −2 ), low‐energy consumption (≈725 aJ), and excellent flexibility with a minimum bending radius of ≈0.85 mm, while enduring 50 000 bending cycles at a 2 mm radius. The OESTs surpass previously reported flexible optoelectronic neuromorphic devices in photosensitivity and minimal bending radius. The OESTs demonstrated capabilities in image memory, extraction, and multi‐level encryption. This work paves the way for future smart windows capable of in‐sensor computing and environmental adaptation.