You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I split PIE and JAAD dataset as your paper and codes. But I find imbalance of C/NC samples.
In PIE dataset, the numbers of different samples are as follows:
NC
C
Train
3576
1194
Test
2742
1074
In JAAD_beh dataset, , the numbers of different samples are as follows:
NC
C
Train
374
1760
Test
704
1177
In PIE dataset, the number of NC samples is far more than the number of C samples. In JAAD dataset, the number of C samples is more than the number of NC samples. I think it is harmful to train a model. Is the split result correct? Could you please explain this distribution?
The text was updated successfully, but these errors were encountered:
Yes, we are aware of the imbalance in the dataset. In JAAD the imbalance is caused by the fact that we were mostly interested in pedestrians who cross. PIE contains continuous driving footage so it corresponds to a more naturalistic distribution of pedestrians, many of whom stand near the road waiting to cross.
Please see the code of the benchmark and the text of the paper for details on how to deal with the imbalance. In short, you can either subsample the larger class or keep the data and set the class weights inversely proportional to the samples in each class. Both methods work well in training, with the second option you get more training data.
Hi,
I split PIE and JAAD dataset as your paper and codes. But I find imbalance of C/NC samples.
In PIE dataset, the numbers of different samples are as follows:
In JAAD_beh dataset, , the numbers of different samples are as follows:
In PIE dataset, the number of NC samples is far more than the number of C samples. In JAAD dataset, the number of C samples is more than the number of NC samples. I think it is harmful to train a model. Is the split result correct? Could you please explain this distribution?
The text was updated successfully, but these errors were encountered: