• Laser & Optoelectronics Progress
  • Vol. 59, Issue 6, 0617026 (2022)
Ying Ji1、*, Lingran Gong1, Shuang Fu2, and Yawei Wang1
Author Affiliations
  • 1School of Physics and Electronic Engineering, Jiangsu University, Zhenjiang , Jiangsu 212013, China
  • 2Department of Biomedical Engineering, Southern University of Science and Technology, Shenzhen , Guangdong 518055, China
  • show less
    DOI: 10.3788/LOP202259.0617026 Cite this Article Set citation alerts
    Ying Ji, Lingran Gong, Shuang Fu, Yawei Wang. Automatic Phase Recognition Method Based on Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2022, 59(6): 0617026 Copy Citation Text show less
    Examples of each class in training dataset (left) and testing dataset (right) (a) Red blood cell; (b) polystyrene bead; (c) small lymphocyte; (d) noise map
    Fig. 1. Examples of each class in training dataset (left) and testing dataset (right) (a) Red blood cell; (b) polystyrene bead; (c) small lymphocyte; (d) noise map
    Structure of CNN implemented for 4-class phase recognition
    Fig. 2. Structure of CNN implemented for 4-class phase recognition
    Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Fig. 3. Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Phase distributions of red blood cell before and after change. (a) Before change; (b) after change
    Fig. 4. Phase distributions of red blood cell before and after change. (a) Before change; (b) after change
    Performance of new CNN model in training process. (a) Loss function value; (b) accuracy
    Fig. 5. Performance of new CNN model in training process. (a) Loss function value; (b) accuracy
    Interferograms of various samples. (a) Red blood cell; (b) polystyrene bead; (c) small lymphocyte; (d) noise map
    Fig. 6. Interferograms of various samples. (a) Red blood cell; (b) polystyrene bead; (c) small lymphocyte; (d) noise map
    Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Fig. 7. Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    ResNet-17 network architecture. (a) Structure of residual block; (b) total architecture of network
    Fig. 8. ResNet-17 network architecture. (a) Structure of residual block; (b) total architecture of network
    Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Fig. 9. Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Polystyrene bead interferograms with different intensity ratios (reference wave∶object wave). (a) 1∶1; (b) 2∶1; (c) 3∶1; (d) 4∶1
    Fig. 10. Polystyrene bead interferograms with different intensity ratios (reference wave∶object wave). (a) 1∶1; (b) 2∶1; (c) 3∶1; (d) 4∶1
    Polystyrene bead interferograms with different fringe spatial frequencies. (a) 1.05 rad/pixel; (b) 1.57 rad/pixel; (c) 2.09 rad/pixel; (d) 3 rad/pixel
    Fig. 11. Polystyrene bead interferograms with different fringe spatial frequencies. (a) 1.05 rad/pixel; (b) 1.57 rad/pixel; (c) 2.09 rad/pixel; (d) 3 rad/pixel
    Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Fig. 12. Performance of LeNet-5 model in training process. (a) Loss function value; (b) accuracy
    Confusion matrixPredicted class
    Red blood cellPolystyrene beadSmall lymphocyteNo phase object
    Actual classRed blood cell2822
    Polystyrene bead482
    Small lymphocyte1535
    No phase object644
    Table 1. Confusion matrix of classification results of LeNet-5 model on interferogram testing dataset
    ClassAccuracy /%Recall /%F1-ScoreOverall accuracy /%
    Red blood cell82.35560.66777.5
    Polystyrene bead76.19960.850
    Small lymphocyte94.59700.805
    No phase object66.67880.759
    Table 2. Performance of LeNet-5 model on each class of interferogram testing dataset
    ClassAccuracy /%Recall /%F1-ScoreOverall accuracy /%
    Red blood cell1001001.00098
    Polystyrene bead100920.958
    Small lymphocyte92.591000.962
    No phase object1001001.000
    Table 3. Performance of ResNet-17 model on each class of interferogram testing dataset
    Datasets with different intensity ratios(RODifferent fringe spatial frequency datasets /(rad⋅pixel-1Correctly identify number of samples/number of all samplesAccuracy /%
    1∶11.05863/867453/86799.5452.25
    2∶11.57849/867863/86797.9299.54
    3∶12.09849/867361/86797.9241.64
    4∶13840/867654/86796.8975.43
    Table 4. Performance of ResNet-17 model on interferogram datasets with different fringe spatial frequencies
    ClassAccuracy /%Recall /%F1-ScoreOverall accuracy /%
    Red blood cell1001001.00097.25
    Polystyrene bead90.091000.948
    Small lymphocyte100890.942
    No phase object1001001.000
    Table 5. Performance of ResNet-17 model on each class of interferogram testing dataset with multiple spatial frequencies
    Ying Ji, Lingran Gong, Shuang Fu, Yawei Wang. Automatic Phase Recognition Method Based on Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2022, 59(6): 0617026
    Download Citation