no code implementations • 8 May 2021 • Yumeng Zhang, Li Chen, Yufeng Liu, Xiaoyan Guo, Wen Zheng, Junhai Yong
Deep learning methods have achieved excellent performance in pose estimation, but the lack of robustness causes the keypoints to change drastically between similar images.
no code implementations • 15 Sep 2020 • Chen Ma, Shuyu Cheng, Li Chen, Jun Zhu, Junhai Yong
In each iteration, SWITCH first tries to update the current sample along the direction of $\hat{\mathbf{g}}$, but considers switching to its opposite direction $-\hat{\mathbf{g}}$ if our algorithm detects that it does not increase the value of the attack objective function.
no code implementations • 16 Sep 2019 • Yumeng Zhang, Gaoguo Jia, Li Chen, Mingrui Zhang, Junhai Yong
The dynamic image compresses the motion information of video into a still image, removing the interference factors such as the background.
no code implementations • 11 Sep 2019 • Yumeng Zhang, Li Chen, Yufeng Liu, Junhai Yong, Wen Zheng
During training, based on the relation between these common characteristics and 3D pose learned from fully-annotated synthetic datasets, it is beneficial for the network to restore the 3D pose of weakly labeled real-world datasets with the aid of 2D annotations and depth images.
1 code implementation • 6 Aug 2019 • Chen Ma, Chenxu Zhao, Hailin Shi, Li Chen, Junhai Yong, Dan Zeng
To solve such few-shot problem with the evolving attack, we propose a meta-learning based robust detection method to detect new adversarial attacks with limited examples.
no code implementations • 29 Mar 2019 • Zihao Bo, Hao Zhang, Junhai Yong, Feng Xu
We propose a real-time DNN-based technique to segment hand and object of interacting motions from depth inputs.
2 code implementations • 14 Dec 2018 • Chen Ma, Li Chen, Junhai Yong
(2) We integrate various dynamic models (including convolutional long short-term memory, two stream network, conditional random field, and temporal action localization network) into AU R-CNN and then investigate and analyze the reason behind the performance of dynamic models.
Ranked #1 on Action Unit Detection on BP4D