For optimization-based methods, we provide configuration files and log files. For learning-based methods, we provide configuration files, log files and pretrained models. Moreover, all supported methods are evaluated on three common benchmarks: Campus, Shelf, and CMU Panoptic.
Please refer to MVPose for details.
Please refer to MVPose with tracking for details.
Please refer to Shape-aware 3D Pose Optimization for details.
Please refer to MvP benchmarks for details.
Please refer to FourDAG benchmarks for details.