• Laser & Optoelectronics Progress
  • Vol. 58, Issue 20, 2015008 (2021)
Hongying Zhang*, Weimin Yang, and Huisan Wang
Author Affiliations
  • College of Electronic Information and Automation, Civil Aviation University of China, Tianjin 300300, China
  • show less
    DOI: 10.3788/LOP202158.2015008 Cite this Article Set citation alerts
    Hongying Zhang, Weimin Yang, Huisan Wang. Face Reconstruction Method Based on Optimized Three-Dimensional Morphable Model Parameters[J]. Laser & Optoelectronics Progress, 2021, 58(20): 2015008 Copy Citation Text show less
    Flow chart of our method
    Fig. 1. Flow chart of our method
    Positive face feature points extracted by ASM. (a) Positive face 1; (b) positive face 2
    Fig. 2. Positive face feature points extracted by ASM. (a) Positive face 1; (b) positive face 2
    Feature points extracted by different methods. (a) ASM; (b) PRN
    Fig. 3. Feature points extracted by different methods. (a) ASM; (b) PRN
    Facial edge features extracted based on Canny operator. (a) Face image; (b) face edge
    Fig. 4. Facial edge features extracted based on Canny operator. (a) Face image; (b) face edge
    Superposition of face edge and face feature points. (a) Face edge; (b) face feature points;(c) superimposed image
    Fig. 5. Superposition of face edge and face feature points. (a) Face edge; (b) face feature points;(c) superimposed image
    Structure of the edge feature detector
    Fig. 6. Structure of the edge feature detector
    Feature points before and after calibration. (a) Before calibration; (b) after calibration
    Fig. 7. Feature points before and after calibration. (a) Before calibration; (b) after calibration
    Projection results. (a) 3D face model; (b) projected image
    Fig. 8. Projection results. (a) 3D face model; (b) projected image
    Result of model rendering. (a) Before rendered; (b) after rendered
    Fig. 9. Result of model rendering. (a) Before rendered; (b) after rendered
    NMSE with different coefficients. (a) λid; (b) λexp
    Fig. 10. NMSE with different coefficients. (a) λid; (b) λexp
    NMSE of reconstructed image by different methods
    Fig. 11. NMSE of reconstructed image by different methods
    3D face reconstruction results. (a) 2D image; (b) 0°; (c) 30°; (d) 60°; (e) 90°
    Fig. 12. 3D face reconstruction results. (a) 2D image; (b) 0°; (c) 30°; (d) 60°; (e) 90°
    Reconstruction results of different methods. (a) 2D image; (b) VRN method; (c) 3DDFA method; (d) 3DMM-CNN method; (e) our method
    Fig. 13. Reconstruction results of different methods. (a) 2D image; (b) VRN method; (c) 3DDFA method; (d) 3DMM-CNN method; (e) our method
    MethodVRN3DDFA3DMM-CNNOurs
    AFW0.6980.6341.2230.617
    IBUG0.6230.6441.0560.545
    LFPW0.6690.6531.3260.656
    Table 1. NMSE of reconstructed image by different methods
    Hongying Zhang, Weimin Yang, Huisan Wang. Face Reconstruction Method Based on Optimized Three-Dimensional Morphable Model Parameters[J]. Laser & Optoelectronics Progress, 2021, 58(20): 2015008
    Download Citation