Deep Learning-Assisted Electro-Thermochromic Fluorescent Fibers for Self-Adaptive Intelligent Display in Dynamic Environments

Electrochromic fibers adept at miniaturization and seamless integration with the soft and dynamically deformed human body, emerge as promising frontrunners in smart displays as a new form of wearable visual output devices. However, it remains a huge challenge to simultaneously achieve conformal smart fibers with various colors, desirable color-changing performance, reliable mechanical properties, and a simple preparation procedure for intelligent display of complex information during a human-machine interface. Herein, full-color electro-thermochromic fluorescent fibers with fast response (<1 s), off–on switching luminescence, good reversibility (>50 cycles), and programmable control, are developed on the basis of self-crystallinity phase changes and Förster resonance energy transfer; and performed self-adaptive intelligent display (≈2 s) of multi-targets in dynamic environments with the assistance of deep learning, by using a wearable textile display system as an example implementation. The work unifies deep learning and smart fibers to make it possible for bio-inspired self-adaptive ability, in which the dynamic targets encompass a diverse array of morphological entities, extending even to stimuli like light, scent, temperature, and humidity. This self-adaptive intelligent display fiber presents unprecedented opportunities for next-generation wearable electronics and systems, fostering seamless communication and adaptability within dynamic environments.