Closed Lucashsmello closed 2 years ago
In the paper, we also discussed the data augmentation, and thus RandomAddGaussian() means randomly adding Gaussian noise in each training sample and is one kind of the data augmentation method. Data augmentation is a common step in image classification, which can sometimes enhance the generalization ability of the model. Thus, we discussed it.
Hi. The file CNN_Datasets/R_NA/datasets/CWRUFFT.py, at line 105, adds noise to the training dataset, despite being used for a classification task. Is that correct?
In the paper, we also discussed the data augmentation, and thus RandomAddGaussian() means randomly adding Gaussian noise in each training sample and is one kind of the data augmentation method. Data augmentation is a common step in image classification, which can sometimes enhance the generalization ability of the model. Thus, we discussed it.
Thanks for your response. That is right. However, shouldn't datasets on R_NA folder be free of data augmentation?
Thanks for your response. That is right. However, shouldn't datasets on R_NA folder be free of data augmentation?
Yes, you are right! We have made a mistake when cleaning up our codes. I have modified this error, which happens in both AE and CNN
Nice! Many thanks. This seems to happen to other datasets in CNN_Datasets/R_NA folder too:
Nice! Many thanks. This seems to happen to other datasets in CNN_Datasets/R_NA folder too:
- CWRUCWT.py
- JNUCWT.py
- MFPTCWT.py
- PUCWT.py
- PU.py
- SEUCWT.py
- UoCCWT.py
- XJTUCWT.py They all use RandomScaler and RandomCrop.
Thanks very much!
Resolved.
Hi. The file CNN_Datasets/R_NA/datasets/CWRUFFT.py, at line 105, adds noise to the training dataset, despite being used for a classification task. Is that correct?