Scholarlink

<p>Autonomous colour correction embedded into an individual pixel is crucial to create next-generation intelligent visual systems. Although existing feedback circuits enable robust ex situ colour correction, they remain bulky with logic complexity. Here we propose in-pixel colour correction by integrating three panchromatic organic active adaptation transistors as a single pixel, each featuring two complementary broadband bulk heterojunctions. The devices display an active adaptation index, that is, a change in photosensitivity as a function of orders of magnitude changes in luminance, of over 150 to red, green and blue light stimuli. More importantly, the subpixels adapt following the von Kries coefficient law, thereby mimicking the ability of a human visual system to adjust to changes in illumination and preserve the appearance of colours. Our proof-of-concept device array, under distorted light conditions, achieves a recognition accuracy of &gt;96.3% in a convolutional neural network simulation. These results represent a key step for constructing a new generation of smart visual systems with in-sensor functionalities.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

In-pixel colour correction with organic self-adaptive transistors

  • 期刊论文

  • 发表日期：

    2026年1月7日

查看全文
  • Zihan He,
  • Wei Wang,
  • Zepang Zhan,
  • Lingxuan Jia,
  • Yutao Ge,
  • Zitong Zhan,
  • Peiyao Xue,
  • Weijie Wang,
  • Lanyi Xiang,
  • Yingqiao Ma,
  • Yawen Li,
  • Zhiyi Li,
  • Xiaojuan Dai,
  • Dekai Ye,
  • Liyao Liu,
  • Fengjiao Zhang,
  • Ye Zou,
  • Yuze Lin,
  • Xiaowei Zhan,
  • Daoben Zhu,
  • Chong-an Di

摘要

Autonomous colour correction embedded into an individual pixel is crucial to create next-generation intelligent visual systems. Although existing feedback circuits enable robust ex situ colour correction, they remain bulky with logic complexity. Here we propose in-pixel colour correction by integrating three panchromatic organic active adaptation transistors as a single pixel, each featuring two complementary broadband bulk heterojunctions. The devices display an active adaptation index, that is, a change in photosensitivity as a function of orders of magnitude changes in luminance, of over 150 to red, green and blue light stimuli. More importantly, the subpixels adapt following the von Kries coefficient law, thereby mimicking the ability of a human visual system to adjust to changes in illumination and preserve the appearance of colours. Our proof-of-concept device array, under distorted light conditions, achieves a recognition accuracy of >96.3% in a convolutional neural network simulation. These results represent a key step for constructing a new generation of smart visual systems with in-sensor functionalities.