• Optical Instruments
  • Vol. 41, Issue 5, 38 (2019)
Huamin WU, Moyu YANG, Xiaoxue HUANG, Caiquan JI..., Weijie WANG, Rongfu ZHANG* and Nan CHEN|Show fewer author(s)
Author Affiliations
  • School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
  • show less
    DOI: 10.3969/j.issn.1005-5630.2019.05.006 Cite this Article
    Huamin WU, Moyu YANG, Xiaoxue HUANG, Caiquan JI, Weijie WANG, Rongfu ZHANG, Nan CHEN. Cross-center detection based on deep learning[J]. Optical Instruments, 2019, 41(5): 38 Copy Citation Text show less
    Neuron model
    Fig. 1. Neuron model
    Convolution operation diagram
    Fig. 2. Convolution operation diagram
    Pooling operation diagram
    Fig. 3. Pooling operation diagram
    Cross image taken far and near the center of curvature radius of the measured mirror
    Fig. 4. Cross image taken far and near the center of curvature radius of the measured mirror
    Cross image of contaminated mirror
    Fig. 5. Cross image of contaminated mirror
    Image preprocessing process
    Fig. 6. Image preprocessing process
    Three ways of annotation
    Fig. 7. Three ways of annotation
    Structural diagram of convolutional neural network
    Fig. 8. Structural diagram of convolutional neural network
    Change of loss value under different marking methods
    Fig. 9. Change of loss value under different marking methods
    Five-fold cross validation to evaluate model performance
    Fig. 10. Five-fold cross validation to evaluate model performance
    Loss of the five-fold cross validation test set
    Fig. 11. Loss of the five-fold cross validation test set
    Key point prediction and center point calculation of four different types
    Fig. 12. Key point prediction and center point calculation of four different types
    Principle of SUSAN algorithm
    Fig. 13. Principle of SUSAN algorithm
    算法清晰十字像边缘不规则十字像模糊十字像对比度低十字像
    标注值(119.65,82.35)(139.82,111.63)(88.85,75.64)(114.35,126.73)
    CNN预测值(118.82,83.41)(138.01,109.73)(88.18,73.83)(111.64,124.91)
    直线拟合(118.96,81.07)(144.69,108.39)(90.19,72.36)无法检测
    SUSAN算法(117,82)(137,112)(84,78)无法检测
    Table 1. Prediction results of cross center coordinates
    Huamin WU, Moyu YANG, Xiaoxue HUANG, Caiquan JI, Weijie WANG, Rongfu ZHANG, Nan CHEN. Cross-center detection based on deep learning[J]. Optical Instruments, 2019, 41(5): 38
    Download Citation