Biomimetic High‐Order Encoder for Trichromatic Recognition Based on Flexible MoS <sub>2</sub> Ring Oscillators
ABSTRACT Biomimetic visual perception and processing have gained increasing prominence in advancing vision restoration technologies and developing next‐generation human‐machine interfaces. As dual fundamental parameters for environmental information capture, light intensity and color perception constitute essential dimensions in visual signal acquisition. While digital encoding remains a conventional approach, the biomimetic frequency coding approach is considered to be an efficient input method for simulating the retina neural architecture in artificial vision systems. Here, we fabricated molybdenum disulfide (MoS 2 ) based ring oscillators for color and intensity resolved image recognition, which exploit wavelength‐dependent photoresponsivity (400–700 nm) to implement biomimetic spectral coding. Through MoS 2 oscillators, light signals are transduced into frequency‐modulated electrical spikes from 265 Hz to 3.8 kHz, establishing hardware foundations for chromatic opponency processing in neuromorphic visual systems. This orthogonal signal modulation strategy enables cross‐talk‐free transmission of visual information, effectively decoupling wavelength‐dependent chromatic data from intensity‐modulated luminance signals. By integrating the obtained oscillatory signals into a convolutional neural network system, we successfully demonstrated the device's applicability in amplifying subtle differences in color and texture features in image recognition and thus realizing high dynamic range encoding and robust feature extraction.
Authors
- Songge Zhang (ORCID: https://orcid.org/0000-0002-8921-5176)
- Shuopei Wang (ORCID: https://orcid.org/0000-0001-7691-0797)
- Xiuzhen Li (ORCID: https://orcid.org/0009-0001-1799-7762)
- Yang Wei (ORCID: https://orcid.org/0000-0001-8162-3901)
- Xingchao Zhang (ORCID: https://orcid.org/0000-0001-9004-8662)
- Luojun Du (ORCID: https://orcid.org/0000-0002-2420-8258)
- Rong Yang (ORCID: https://orcid.org/0000-0003-2216-7584)
- Yixiao Li (ORCID: https://orcid.org/0009-0001-9140-9834)
- Xiaoming Tao (ORCID: https://orcid.org/0000-0002-2406-0695)
- Na Li (ORCID: https://orcid.org/0000-0002-0392-6672)
- Yuchen Wang (ORCID: https://orcid.org/0000-0001-9194-7109)
- Guangyu Zhang (ORCID: https://orcid.org/0000-0002-1242-4391)
- Hua Yu (ORCID: https://orcid.org/0000-0003-4610-932X)
- Yuchao Zhou (ORCID: https://orcid.org/0000-0002-9665-4206)
- Dongxia Shi
- Chenyang Cui (ORCID: https://orcid.org/0009-0007-3672-400X)
- Su Liu
- Y. Q. Zhang (ORCID: https://orcid.org/0009-0001-2407-1204)
Institutions
- Hong Kong Polytechnic University (HK)
- Changsha University (CN)
- Songshan Lake Materials Laboratory (CN)
- Beijing National Laboratory for Molecular Sciences (CN)
- FZU ‒ Institute of Physics of the Academy of Sciences of the Czech Republic (CZ)
- University of Chinese Academy of Sciences (CN)
Publication Details
- Journal
- Advanced Electronic Materials
- Published
- 2025-11-26
- DOI
- https://doi.org/10.1002/aelm.202500486
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00