sajjadmzf / pedestrian_datasets_vis

The visualization of annotation files for different pedestrian datasets.

Pedestrian Datasets Visualization

The visualization of annotation files for different pedestrian datasets.

Stanford Drone Dataset

You can find the dataset here(http://cvgl.stanford.edu/projects/uav_data/).

Their Publication: A. Robicquet, A. Sadeghian, A. Alahi, S. Savarese, Learning Social Etiquette: Human Trajectory Prediction In Crowded Scenes in European Conference on Computer Vision (ECCV), 2016.

Examples of their annotations (As you can see, they are prone to annotation errors):

ETH/UCY Datasets:

The video files of these dataset aren't published and the annotations are normalized to (0,1)

Examples of the annotations:

Note: UCY data are interpolated in missing frames.

pedestrian_datasets_vis's Issues

Question about the ETH/UCY data normalization

Dear Author

Thanks for your sharing. May I ask: what is the data normalization procedure for the ETH/UCY dataset?

I notice that the original txt file for those two datasets contains negative coordinates. Therefore I am wondering how did you normalize them.

Thanks