Table 3 Convolutional neural network model settings for combined inputs (sinogram and CT image)
From: Improved computer-aided detection of pulmonary nodules via deep learning in the sinogram domain
Layer | Parameters |
|---|---|
L1_1 (sinogram) | Conv, 7 × 7, 32, LeakyReLU |
L2_1 (sinogram) | Maxpooling, 2 × 2, stride 2 |
L3_1 (sinogram) | Conv, 5 × 5, 64, LeakyReLU |
L4_1 (sinogram) | Maxpooling, 2 × 2, stride 2 |
L1_2 (CT image) | Conv, 7 × 7, 64, LeakyReLU |
L2_2 (CT image) | Maxpooling, 2 × 2, stride 2 |
L3_2 (CT image) | Conv, 5 × 5, 64, LeakyReLU |
L4_2 (CT image) | Maxpooling, 2 × 2, stride 2 |
L5 | Fully-Connected,1000, LeakyReLU |
L6 | Fully-Connected,2, Softmax |