Skip to content

Some additions, optimizations and modifications to the source code base based on the VIBE paper reproduction

License

Notifications You must be signed in to change notification settings

Xushuolin/VIBE_XU_changed

Repository files navigation

VIBE_XU_changed

This paper has very many bugs in its reproduction, especially in Win systems. It is therefore highly recommended to deploy it step by step in LINUX with reference to the original installation requirements, in particular, you should open the requirements file and install the requirements library step by step. However, after recent maintenance by the author, it should be much better. The last point to note is about Pyglt, which should be installed around version 1.5.0 for this task to work properly. Some additions, optimizations and modifications to the source code base based on the VIBE paper reproduction

Mainly modified the output type in the demo to show the predicted 3D joints.

On the other hand, to achieve this effect, a new renderer was rewritten in Render and called in the demo.

https://livebournemouthac-my.sharepoint.com/:v:/g/personal/s5529054_bournemouth_ac_uk/EeAguu3ZmZ9GhVDVSomgGpgB0YQXxMlNfgCRMDRvwcDUUA?e=TLg76c

According kocabas (2020) to did

References

Kocabas, Muhammed and Athanasiou, Nikos and Black, Michael J.. (2020). VIBE: Video Inference for Human Body Pose and Shape Estimation. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 10(1), 1-20.

About

Some additions, optimizations and modifications to the source code base based on the VIBE paper reproduction

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published