IndraLab

Statements

ResNet18 binds CALM1. 5 / 5
| 5
sparser
"As shown, after dimensionality reduction and visualization, the four fault types are well distinguished in the CAM-ResNet18 output."
40594798
sparser
"Figure 12 presents the confusion matrix for fault recognition using CIMAM-ResNet18 and CAM-ResNet18."
40594798
sparser
"The matrix reveals significant confusion between faults gdt and nq in CAM-ResNet18, with a fault recognition accuracy of 95.78%."
40594798
sparser
"The preprocessed data were classified using the CAM-ResNet18 and CIMAM-ResNet18 models, and the high-dimensional features were visualized and reduced using T-SNE, as shown in Fig. 10."
40594798
sparser
"As shown in Fig.  xref , the loss function curve of the CAM-ResNet18 model exhibits a convergence trend, but fluctuations persist during the convergence phase, leading to overfitting and affecting fault recognition accuracy."
40594798