• Acta Optica Sinica
  • Vol. 40, Issue 22, 2210003 (2020)
Tao Zhang1、2, Qin Zeng1、2、*, Wenli Du1、2, and Hao Wang1、2
Author Affiliations
  • 1School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China
  • 2Texas Instruments DSP Joint Lab, Tianjin University, Tianjin 300072, China
  • show less
    DOI: 10.3788/AOS202040.2210003 Cite this Article Set citation alerts
    Tao Zhang, Qin Zeng, Wenli Du, Hao Wang. Regional Backlight Brightness Extraction Algorithm Based on Deep Learning[J]. Acta Optica Sinica, 2020, 40(22): 2210003 Copy Citation Text show less
    Data measurement method based on local backlight extraction and diagram of system structure
    Fig. 1. Data measurement method based on local backlight extraction and diagram of system structure
    Flowcharts of dividing methods. (a) Mean plus/minus variance method; (b) percentage dividing method
    Fig. 2. Flowcharts of dividing methods. (a) Mean plus/minus variance method; (b) percentage dividing method
    Measurement system of local backlight extraction
    Fig. 3. Measurement system of local backlight extraction
    Test environment of local backlight extraction. (a) Experimental setup; (b) dark environment
    Fig. 4. Test environment of local backlight extraction. (a) Experimental setup; (b) dark environment
    Diagram of local dimming system based on deep learning
    Fig. 5. Diagram of local dimming system based on deep learning
    Proposed BENN structure
    Fig. 6. Proposed BENN structure
    Structure of Res module
    Fig. 7. Structure of Res module
    Simulation results. (a)(c)(e)(g) Traditional algorithm; (b)(d)(f)(h) proposed algorithm
    Fig. 8. Simulation results. (a)(c)(e)(g) Traditional algorithm; (b)(d)(f)(h) proposed algorithm
    Results of display system. (a)(c)(e)(g)(i)(k)(m)(o)(q)(s)(u)(w) Traditional algorithm; (b)(d)(f)(h)(j)(l)(n)(p)(r)(t)(v)(x) proposed algorithm
    Fig. 9. Results of display system. (a)(c)(e)(g)(i)(k)(m)(o)(q)(s)(u)(w) Traditional algorithm; (b)(d)(f)(h)(j)(l)(n)(p)(r)(t)(v)(x) proposed algorithm
    AlgorithmPSNRSSIMCR
    Proposed method27.96720.97932.9293
    Otsu27.24070.8982.2408
    LUT27.21520.89582.2732
    CDF27.19570.88672.1883
    Average method27.23770.85382.0705
    Table 1. Objective valuation indexes
    MethodPSNRSSIMCR
    BENN27.90720.97932.9293
    BENN-5×5 Conv27.91950.97582.9230
    BENN-R_127.90630.97952.9187
    BENN-5×5 Conv-R_127.73750.96852.9228
    Table 2. Experimental results of different networks
    Tao Zhang, Qin Zeng, Wenli Du, Hao Wang. Regional Backlight Brightness Extraction Algorithm Based on Deep Learning[J]. Acta Optica Sinica, 2020, 40(22): 2210003
    Download Citation