Hon-Wong / PTSEFormer

[ECCV 2022] PTSEFormer: Progressive Temporal-Spatial Enhanced TransFormer Towards Video Object Detection
https://arxiv.org/abs/2209.02242
MIT License
34 stars 7 forks source link

Imagesets txt files #13

Open iamamiramine opened 11 months ago

iamamiramine commented 11 months ago

Can you help me with how to organize the Imageset text files? I understand for each video I should have one text file, and in this text file is the path of the video frames and their frame id. Am I missing something here? Because I organized the content of the text files in this way:

frame_title frame_index

and modified the vidmulti.py script accordingly, but I am facing a different error (I posted it in issue #12) Any help on how to organize the Imagesets text files? Thank you