• Laser & Optoelectronics Progress
  • Vol. 62, Issue 5, 0530001 (2025)
Jiajia Wang1,*, Qianqian Mo1, and Tao Yang1,2
Author Affiliations
  • 1Institute of Advanced Materials, Nanjing University of Posts and Telecommunications, Nanjing 210023, Jiangsu , China
  • 2Henan Institute of Flexible Electronics, Zhengzhou 450046, Henan , China
  • show less
    DOI: 10.3788/LOP241318 Cite this Article Set citation alerts
    Jiajia Wang, Qianqian Mo, Tao Yang. Deep Learning-Based Disordered-Dispersion Miniature Spectrometer[J]. Laser & Optoelectronics Progress, 2025, 62(5): 0530001 Copy Citation Text show less
    DL-based disordered-dispersion miniature spectrometer
    Fig. 1. DL-based disordered-dispersion miniature spectrometer
    Rough surface of frosted glass
    Fig. 2. Rough surface of frosted glass
    Speckle intensity under different wavelength monochromatic light illumination. (a) 514 nm; (b) 516 nm; (c) their normalized intensity difference
    Fig. 3. Speckle intensity under different wavelength monochromatic light illumination. (a) 514 nm; (b) 516 nm; (c) their normalized intensity difference
    Normalized spectra of 8 LEDs
    Fig. 4. Normalized spectra of 8 LEDs
    MLP neural network framework
    Fig. 5. MLP neural network framework
    Loss function (MSE) decline curve with increasing training epochs
    Fig. 6. Loss function (MSE) decline curve with increasing training epochs
    Spectral reconstruction results of the spectrometer. (a) Reconstruction of 6 narrowband spectra with different wavelength; (b) reconstruction of 2 broadband spectra at different times; (c) reconstruction of 2 close narrowband spectra; (d) comparison of spectral reconstruction using Tikhonov regularization and MLP neural network
    Fig. 7. Spectral reconstruction results of the spectrometer. (a) Reconstruction of 6 narrowband spectra with different wavelength; (b) reconstruction of 2 broadband spectra at different times; (c) reconstruction of 2 close narrowband spectra; (d) comparison of spectral reconstruction using Tikhonov regularization and MLP neural network