EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Trap‐Mediated Programmable Adaptive Plasticity in a Reconfigurable Neuromorphic Transistor for Retina‐to‐Cortex‐Inspired Visual Processing

Xiaomin Wu, Wenxin Zhang, Yunjian Liang, Di Liu, Minrui Lian, Shuiyuan Chen, Huipeng Chen

SmartMat · 2026-10 · 7 (5)

DOI: 10.1002/smm2.70117 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SmartMat, online erschienen am 15.10.2026. Verfasst von Xiaomin Wu und 6 weiteren Autor:innen. Die Arbeit nennt 56 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Natural Science Foundation of Fujian Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dynamic machine vision requires compact neuromorphic hardware capable of adaptive sensing and robust information processing under fluctuating illumination. However, most neuromorphic visual systems still rely on separated modules for electrical switching, retina‐like preprocessing, synaptic memory, cortex‐inspired computation, limiting compactness and reconfigurability. Here, we report a trap‐engineered reconfigurable neuromorphic transistor based on a hybrid organic indacenodithiophene‐benzothiadiazole/polymer electret heterostructure for integrated retina‐to‐cortex‐inspired visual processing. Electrical gate pulses drive synaptic operation and information processing, while background illumination and gate bias jointly program the adaptive state through shallow/deep‐trap occupation and carrier capture/release. By regulating these material‐level trap dynamics, the device integrates transistor‐like switching, retina‐inspired adaptive preprocessing, and cortex‐inspired synaptic memory and computation within a single architecture. The transistor exhibits rapid visual adaptation within 0.39 s and a broad light‐intensity regulation range exceeding 73 dB. Enabled by programmable adaptive plasticity, it achieves robust contour extraction and visual recognition under strong illumination variations, including semantic segmentation accuracy above 97.3% and short‐term traffic trajectory prediction accuracy of 93.96% under dynamic exposure. This work links charge‐trapping physics with adaptive neuromorphic vision, offering a compact strategy for resilient machine vision hardware.

Förderung

  • National Natural Science Foundation of China
  • Natural Science Foundation of Fujian Province

Zitieren

APA

Wu, X., Zhang, W., Liang, Y., Liu, D., Lian, M., Chen, S., & Chen, H. (2026). Trap‐Mediated Programmable Adaptive Plasticity in a Reconfigurable Neuromorphic Transistor for Retina‐to‐Cortex‐Inspired Visual Processing. SmartMat, 7(5). https://doi.org/10.1002/smm2.70117

Vancouver

Wu X, Zhang W, Liang Y, Liu D, Lian M, Chen S, et al. Trap‐Mediated Programmable Adaptive Plasticity in a Reconfigurable Neuromorphic Transistor for Retina‐to‐Cortex‐Inspired Visual Processing. SmartMat. 2026;7(5). doi:10.1002/smm2.70117

BibTeX
@article{Wu2026,
  author = {Wu, Xiaomin and Zhang, Wenxin and Liang, Yunjian and Liu, Di and Lian, Minrui and Chen, Shuiyuan and Chen, Huipeng},
  title = {Trap‐Mediated Programmable Adaptive Plasticity in a Reconfigurable Neuromorphic Transistor for Retina‐to‐Cortex‐Inspired Visual Processing},
  journal = {SmartMat},
  year = {2026},
  volume = {7},
  number = {5},
  doi = {10.1002/smm2.70117}
}
Kostenlos starten →