Table 2 Details of data augmentation during training
From: Development and performance evaluation of a deep learning lung nodule detection system
Data augmentation type | Method |
|---|---|
Rotation | Randomly rotate around the z-axis in the range of − 10 to 10 degrees |
Scale | Randomly change the size by − 15 to 15% in each axis direction |
Sharpness | Change the sharpness according to the following formula |
\(I_{out} = I_{in} + \alpha \left( {I_{in} - f\left( {I_{in} } \right)} \right)\) | |
Iin: input image; Iout: output image; f(Iin): image processed with a Gaussian filter with a standard deviation of 3.0; α: parameter (randomly selected within the range of 0.0–5.0 | |
Smoothing | Process with a Gaussian filter that randomly sets the standard deviation in the range of 0.5–2.0 |
Gaussian Noise | Add Gaussian noise generated in the range of standard deviation 0.0–0.2 |