3D Point Cloud GAN from Human Action Images

Research project, SIMSLAB Laboratory, NTUST, 2020. With Hendrik Tampubolon.

Goal: build an action-recognition model based on 3D point clouds derived from still-image data, and a generative adversarial network to produce that point-cloud data.

Code on GitHub

References

• Shu, Park, Kwon. “3D point cloud generative adversarial network based on tree structured graph convolutions.” ICCV, 2019.
• Safaei, Balouchian, Foroosh. “UCF-STAR: A Large Scale Still Image Dataset for Understanding Human Actions.” AAAI, 2020.
• Kanazawa et al. “End-to-end recovery of human shape and pose.” CVPR, 2018.
• Wang et al. “Mining actionlet ensemble for action recognition with depth cameras.” CVPR, 2012.