Feature boosting network for 3D pose estimation

In this paper, a feature boosting network is proposed for estimating 3D hand pose and 3D body pose from a single RGB image. In this method, the features learned by the convolutional layers are boosted with a new long short-term dependence-aware (LSTD) module, which enables the intermediate convoluti...

Full description

Saved in:
Bibliographic Details
Main Authors: Liu, Jun, Ding, Henghui, Shahroudy, Amir, Duan, Ling-Yu, Jiang, Xudong, Wang, Gang, Kot, Alex C.
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2022
Subjects:
Online Access:https://hdl.handle.net/10356/154881
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-154881
record_format dspace
spelling sg-ntu-dr.10356-1548812022-01-13T01:49:12Z Feature boosting network for 3D pose estimation Liu, Jun Ding, Henghui Shahroudy, Amir Duan, Ling-Yu Jiang, Xudong Wang, Gang Kot, Alex C. School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering 3D Pose Estimation Convolutional LSTM In this paper, a feature boosting network is proposed for estimating 3D hand pose and 3D body pose from a single RGB image. In this method, the features learned by the convolutional layers are boosted with a new long short-term dependence-aware (LSTD) module, which enables the intermediate convolutional feature maps to perceive the graphical long short-term dependency among different hand (or body) parts using the designed Graphical ConvLSTM. Learning a set of features that are reliable and discriminatively representative of the pose of a hand (or body) part is difficult due to the ambiguities, texture and illumination variation, and self-occlusion in the real application of 3D pose estimation. To improve the reliability of the features for representing each body part and enhance the LSTD module, we further introduce a context consistency gate (CCG) in this paper, with which the convolutional feature maps are modulated according to their consistency with the context representations. We evaluate the proposed method on challenging benchmark datasets for 3D hand pose estimation and 3D full body pose estimation. Experimental results show the effectiveness of our method that achieves state-of-the-art performance on both of the tasks. National Research Foundation (NRF) ROSE is supported by the National Research Foundation, Singapore, under the IDM Strategic Research Programme. This work was in part supported by the National Basic Research Program of China under grant 2015CB351806, and the National Natural Science Foundation of China under Grant 61661146005 and Grant U1611461. We thank NVIDIA AI Technology Centre for GPU donation. 2022-01-13T01:49:12Z 2022-01-13T01:49:12Z 2020 Journal Article Liu, J., Ding, H., Shahroudy, A., Duan, L., Jiang, X., Wang, G. & Kot, A. C. (2020). Feature boosting network for 3D pose estimation. IEEE Transactions On Pattern Analysis and Machine Intelligence, 42(2), 494-501. https://dx.doi.org/10.1109/TPAMI.2019.2894422 0162-8828 https://hdl.handle.net/10356/154881 10.1109/TPAMI.2019.2894422 30676946 2-s2.0-85077941122 2 42 494 501 en IEEE Transactions on Pattern Analysis and Machine Intelligence © 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Engineering::Electrical and electronic engineering
3D Pose Estimation
Convolutional LSTM
spellingShingle Engineering::Electrical and electronic engineering
3D Pose Estimation
Convolutional LSTM
Liu, Jun
Ding, Henghui
Shahroudy, Amir
Duan, Ling-Yu
Jiang, Xudong
Wang, Gang
Kot, Alex C.
Feature boosting network for 3D pose estimation
description In this paper, a feature boosting network is proposed for estimating 3D hand pose and 3D body pose from a single RGB image. In this method, the features learned by the convolutional layers are boosted with a new long short-term dependence-aware (LSTD) module, which enables the intermediate convolutional feature maps to perceive the graphical long short-term dependency among different hand (or body) parts using the designed Graphical ConvLSTM. Learning a set of features that are reliable and discriminatively representative of the pose of a hand (or body) part is difficult due to the ambiguities, texture and illumination variation, and self-occlusion in the real application of 3D pose estimation. To improve the reliability of the features for representing each body part and enhance the LSTD module, we further introduce a context consistency gate (CCG) in this paper, with which the convolutional feature maps are modulated according to their consistency with the context representations. We evaluate the proposed method on challenging benchmark datasets for 3D hand pose estimation and 3D full body pose estimation. Experimental results show the effectiveness of our method that achieves state-of-the-art performance on both of the tasks.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Liu, Jun
Ding, Henghui
Shahroudy, Amir
Duan, Ling-Yu
Jiang, Xudong
Wang, Gang
Kot, Alex C.
format Article
author Liu, Jun
Ding, Henghui
Shahroudy, Amir
Duan, Ling-Yu
Jiang, Xudong
Wang, Gang
Kot, Alex C.
author_sort Liu, Jun
title Feature boosting network for 3D pose estimation
title_short Feature boosting network for 3D pose estimation
title_full Feature boosting network for 3D pose estimation
title_fullStr Feature boosting network for 3D pose estimation
title_full_unstemmed Feature boosting network for 3D pose estimation
title_sort feature boosting network for 3d pose estimation
publishDate 2022
url https://hdl.handle.net/10356/154881
_version_ 1722355334688800768