What's the mean of loss_mask_list and non_linear_ped ? More explain about dataset

agrimgupta92 / sgan

Code for "Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks", Gupta et al, CVPR 2018

MIT License

813 stars 261 forks source link

What's the mean of loss_mask_list and non_linear_ped ? More explain about dataset #66

Open LittleFlyFish opened 4 years ago

LittleFlyFish commented 4 years ago

I have see the data useage. But I don't understand what line means in the data.txt. Could you explain more about the data in the txt file?

And I see these lines: obs_traj, pred_traj, obs_traj_rel, pred_traj_rel, non_linear_ped,
loss_mask, seq_start_end
I know about obs_traj and pred_traj, but how about the rest numbers? What's these variables means?

I need to do my research based on these code. So I am really appreciate if you can give more details of the data. Thank you very much! ^_^

VishalSowrirajan commented 3 years ago

@LittleFlyFish obs_traj_rel and pred_traj_rel are the frame-wise coordinate difference for each pedestrian seq_start_end is used to identify the pedestrians in different sequences. For instance, in sequence 1, we might have 3 pedestrians and the next sequence might constitute 5 pedestrians. To distinguish, seq_start_end variable is used. Loss_mask - I am not sure