EVAD
EVAD copied to clipboard
[ICCV 2023] Efficient Video Action Detection with Token Dropout and Context Refinement
Hi author, I am very interested in your article, but would like to reproduce your code with the jhmdb dataset. Can you provide some code for the processing of the...
Hi! Thanks for your work. I have been trying to run the evaluation script. However, I encountered an issue which I don't seem to have the knowledge to solve it....
非常感谢作者提供的视频动作检测源码,请问有计划提供演示demo吗?关于演示demo,有几个问题需要咨询一下作者: 1:是否和端到端的yowov2动作检测一样,只需要输入一段视频clip即可,但看提供的projects/evad/test_net.py 还需要输入人的坐标位置等信息,请问在实际用的时候,这个坐标位置信息是否可以填写为none,而人的坐标位置信息仅用于测试模型map指标?
Could you provide the complete code about ucf and jhmdb?