More Training data for R-CNN detector causes overfitting?

1 回表示 (過去 30 日間)
Abdussalam Elhanashi
Abdussalam Elhanashi 2019 年 10 月 10 日
回答済み: Abdussalam Elhanashi 2019 年 10 月 16 日
Hi Guys
I am experiencing that when i am using a R-CNN detector for object detection , when i increase the training data , i have bad classification and overfitting
Best,

採用された回答

Prabhan Purwar
Prabhan Purwar 2019 年 10 月 16 日
編集済み: Prabhan Purwar 2019 年 10 月 16 日
Hi,
Overfitting happens when the model fits too well to the training set. It then becomes difficult for the model to generalize to new examples that were not in the training set. For example, model recognizes specific images in the training set instead of general patterns. Training accuracy will be higher than the accuracy on the validation/test set.
Steps for reducing overfitting:
  • Add more variant Dataset
  • Make use of balance Dataset
  • Use data augmentation
  • Use architectures that generalize well
  • Add regularization (mostly dropout, L1/L2 regularization are also possible)
Refer to the following link for further information:

その他の回答 (1 件)

Abdussalam Elhanashi
Abdussalam Elhanashi 2019 年 10 月 16 日
Thanks Prabhan

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by