search

JacobYuan7 / DIN-Group-Activity-Recognition-Benchmark

[ICCV 2021] A new codebase containing various methods for Group Activity Recognition. Paper title: Spatio-Temporal Dynamic Inference Network for Group Activity Recognition.
MIT License
52 stars 15 forks source link

About the visualization #17

Closed SatMa34 closed 1 year ago

SatMa34 commented 1 year ago

Hi Jacob, thanks for your great job! You draw the heatmaps of group interaction graph and person interaction graph in the paper (figure 5), and i notice that in dynamic_infer_module.py, you also write code to save relation_list for visualize, i wonder how your draw the heatmaps, could you share the scripts?

JacobYuan7 commented 1 year ago

Hi Jacob, thanks for your great job! You draw the heatmaps of group interaction graph and person interaction graph in the paper (figure 5), and i notice that in dynamic_infer_module.py, you also write code to save relation_list for visualize, i wonder how your draw the heatmaps, could you share the scripts?

Actually, it is very intuitive. You can try transforming the value to a discrete distribution and try scaling the value to make it distinguishable in the image.

JacobYuan7 commented 1 year ago

I am closing this issue. Feel free to reopen it if you have further questions.