Adaptive Fisher-Based Deep Convolutional Neural Network and Its Application to Recognition of Rolling Element Bearing Fault Patterns and Sizes

<div>(a) Diagnostic classification of DCNN (diagnosis accuracy 82.5%). (b) Diagnostic classification of FDCNN (diagnosis accuracy 92.5%).</div>

Mathematical Problems in Engineering

fig4

Figure 4

Figure 4: Adaptive Fisher-Based Deep Convolutional Neural Network and Its Application to Recognition of Rolling Element Bearing Fault Patterns and Sizes 

Figure 4 | Adaptive Fisher-Based Deep Convolutional Neural Network and Its Application to Recognition of Rolling Element Bearing Fault Patterns and Sizes 