MotionBERT: A Unified Pretraining Framework for Human Motion Analysis

Home

Discover

Bookmarks

Keep Up With Latest Trending Papers. Computer Science, AI and Machine Learning and more.Subscribe

CC BY-SA

Source arXiv

Computer Vision

MotionBERT: A Unified Pretraining Framework for Human Motion Analysis

MotionBERT: Unified Pretraining for Human Motion Analysis

We present a unified pretraining framework to tackle different sub-tasks of human motion analysis including 3d pose estimation, skeleton-based action recognition, and mesh recovery.

The proposed framework is capable of utilizing all kinds of human motion data resources, including motion capture data and in-the-wild videos.

It could capture long-range spatio-temporal relationships among the joints comprehensively and adaptively, exemplified by the lowest 3d pose estimation error so far when trained from scratch.

The proposed framework achieves state-of-the-art performance on all three downstream tasks by simply finetuning the pretrained motion encoder with 1-2linear layers, which demonstrates the versatility of the learned motionrepresentations.

Authors

Wentao Zhu, Xiaoxuan Ma, Zhaoyang Liu, Libin Liu, Wayne Wu, Yizhou Wang

◐ Recommended Follows

◐ Latest Activity