出品 | 深度学习这件小事公众号如需转载，请联系后台授权 计算机视觉（1月8日更新版） [1] TrackFormer: Multi-Object Tracking with Transformers作者 | Tim Meinhardt, Alexander Kirillov, Laura Leal-Taixe, Christoph Feichtenhofer链接 | https://arxiv.org/abs/2101.02702 备注 | Tech. report[2] PVA: Pixel-aligned Volumetric Avatars作者 | Amit Raj, Michael Zollhoefer, Tomas Simon, Jason Saragih, Shunsuke Saito, James Hays, Stephen Lombardi链接 | https://arxiv.org/abs/2101.02697 项目链接 | https://volumetric-avatars.github.io/[3] Where2Act: From Pixels to Actions for Articulated 3D Objects作者 | Kaichun Mo, Leonidas Guibas, Mustafa Mukadam, Abhinav Gupta, Shubham Tulsiani链接 | https://arxiv.org/abs/2101.02692 [4] Self-Supervised Pretraining of 3D Features on any Point-Cloud作者 | Zaiwei Zhang, Rohit Girdhar, Armand Joulin, Ishan M
………………………………
