• Acta Optica Sinica
  • Vol. 39, Issue 6, 0615006 (2019)
Yu Feng, Benshun Yi*, Chenyue Wu, and Yungang Zhang
Author Affiliations
  • Electronic Information School, Wuhan University, Wuhan, Hubei 430072, China
  • show less
    DOI: 10.3788/AOS201939.0615006 Cite this Article Set citation alerts
    Yu Feng, Benshun Yi, Chenyue Wu, Yungang Zhang. Pulmonary Nodule Recognition Based on Three-Dimensional Convolution Neural Network[J]. Acta Optica Sinica, 2019, 39(6): 0615006 Copy Citation Text show less
    2D and 3D convolutions. (a) 2D convolution; (b) 3D convolution
    Fig. 1. 2D and 3D convolutions. (a) 2D convolution; (b) 3D convolution
    Structure of SE block
    Fig. 2. Structure of SE block
    Schematic of SE-Dense block
    Fig. 3. Schematic of SE-Dense block
    Model of network structure
    Fig. 4. Model of network structure
    Recognition results of candidate nodules. (a) Prediction probability of true nodule; (b) prediction probability of pseudopositive nodule
    Fig. 5. Recognition results of candidate nodules. (a) Prediction probability of true nodule; (b) prediction probability of pseudopositive nodule
    MethodFalse positives per scanCPM
    1/81/41/21248
    Model_10.6290.7350.8070.8650.9010.9170.9280.826
    Model_20.7340.8120.8690.9010.9180.9270.9290.870
    Model_30.7540.8210.8880.9130.9300.9330.9370.883
    Proposed0.8070.8430.8770.9110.9250.9340.9390.891
    Table 1. Comparison of pulmonary nodule recognition performance by different network structures on LUNA16 dataset
    AlgorithmFalse positives per scanCPM
    0.1250.250.51248
    Ref. [6]0.6780.7380.8160.8480.8790.9070.9220.827
    Ref. [5]0.6920.7100.8090.8630.8950.9140.9230.838
    Ref. [18]0.7600.7940.8330.8600.8760.8930.9060.846
    Ref. [7]0.8020.8470.8860.9090.9250.9360.9410.892
    Proposed0.8070.8430.8770.9110.9250.9340.9390.891
    Table 2. Pulmonary nodule recognition performance by different algorithms on LUNA16
    Yu Feng, Benshun Yi, Chenyue Wu, Yungang Zhang. Pulmonary Nodule Recognition Based on Three-Dimensional Convolution Neural Network[J]. Acta Optica Sinica, 2019, 39(6): 0615006
    Download Citation