Towards Fast and Accurate Multi-Person Pose Estimation on Mobile Devices
Abstract
The rapid development of autonomous driving, abnormal behavior detection, and behavior recognition makes an increasing demand for multi-person pose estimation-based applications, especially on mobile platforms. However, to achieve high accuracy, state-of-the-art methods tend to have a large model size and complex post-processing algorithm, which costs intense computation and long end-to-end latency. To solve this problem, we propose an architecture optimization and weight pruning framework to accelerate inference of multi-person pose estimation on mobile devices. With our optimization framework, we achieve up to 2.51X faster model inference speed with higher accuracy compared to representative lightweight multi-person pose estimator.
Cite
Text
Shen et al. "Towards Fast and Accurate Multi-Person Pose Estimation on Mobile Devices." International Joint Conference on Artificial Intelligence, 2021. doi:10.24963/IJCAI.2021/715Markdown
[Shen et al. "Towards Fast and Accurate Multi-Person Pose Estimation on Mobile Devices." International Joint Conference on Artificial Intelligence, 2021.](https://mlanthology.org/ijcai/2021/shen2021ijcai-fast/) doi:10.24963/IJCAI.2021/715BibTeX
@inproceedings{shen2021ijcai-fast,
title = {{Towards Fast and Accurate Multi-Person Pose Estimation on Mobile Devices}},
author = {Shen, Xuan and Yuan, Geng and Niu, Wei and Ma, Xiaolong and Guan, Jiexiong and Li, Zhengang and Ren, Bin and Wang, Yanzhi},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2021},
pages = {5012-5015},
doi = {10.24963/IJCAI.2021/715},
url = {https://mlanthology.org/ijcai/2021/shen2021ijcai-fast/}
}