About Augment Images for Training

Question

DONGHYUN KIM 2019-9-2

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/478506-about-augment-images-for-training

回答： Mahesh Taparia 2019-9-20

We know that new augmentation methods are applied randomly at all times.

Insufficient learning, the result is slightly different even though it's the same epoch due to the same data. Can you tell me if this is due to insufficient learning?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Mahesh Taparia 2019-9-20

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/478506-about-augment-images-for-training#answer_392731

Hi,

You can try with all possible data augmentation techniques like flip, rotation, cropping, scaling, adding noise, translation etc. If it won’t work, there may be the chance that your dataset became invariant even after doing data augmentation due to enough data.