Hough forest with optimized leaves for global hand pose estimation with arbitrary postures
Vision-based hand pose estimation is important in human-computer interaction. While many recent works focus on full degree-of-freedom hand pose estimation, robust estimation of global hand pose remains a challenging problem. This paper presents a novel algorithm to optimize the leaf weights in a Hou...
Saved in:
Main Authors: | , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2020
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/139907 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-139907 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1399072020-05-22T07:58:50Z Hough forest with optimized leaves for global hand pose estimation with arbitrary postures Liang, Hui Yuan, Junsong Lee, Jun Ge, Liuhao Thalmann, Daniel School of Electrical and Electronic Engineering Institute for Media Innovation (IMI) Engineering::Electrical and electronic engineering Gesture Recognition Hand Pose Estimation Vision-based hand pose estimation is important in human-computer interaction. While many recent works focus on full degree-of-freedom hand pose estimation, robust estimation of global hand pose remains a challenging problem. This paper presents a novel algorithm to optimize the leaf weights in a Hough forest to assist global hand pose estimation with a single depth camera. Different from traditional Hough forest, we propose to learn the vote weights stored at the leaf nodes of a forest in a principled way to minimize average pose prediction error, so that ambiguous votes are largely suppressed during prediction fusion. Experiments show that the proposed method largely improves pose estimation accuracy with optimized leaf weights on both synthesis and real datasets and performs favorably compared to state-of-the-art convolutional neural network-based methods. On real-world depth videos, the proposed method demonstrates improved robustness compared to several other recent hand tracking systems from both industry and academy. Moreover, we utilize the proposed method to build virtual/augmented reality applications to allow users to manipulate and examine virtual objects with bare hands. MOE (Min. of Education, S’pore) 2020-05-22T07:58:50Z 2020-05-22T07:58:50Z 2017 Journal Article Liang, H., Yuan, J., Lee, J., Ge, L., & Thalmann, D. (2019). Hough forest with optimized leaves for global hand pose estimation with arbitrary postures. IEEE Transactions on Cybernetics, 49(2), 527-541. doi:10.1109/TCYB.2017.2779800 2168-2267 https://hdl.handle.net/10356/139907 10.1109/TCYB.2017.2779800 29990273 2 49 527 541 en IEEE Transactions on Cybernetics © 2017 IEEE. All rights reserved. |
institution |
Nanyang Technological University |
building |
NTU Library |
country |
Singapore |
collection |
DR-NTU |
language |
English |
topic |
Engineering::Electrical and electronic engineering Gesture Recognition Hand Pose Estimation |
spellingShingle |
Engineering::Electrical and electronic engineering Gesture Recognition Hand Pose Estimation Liang, Hui Yuan, Junsong Lee, Jun Ge, Liuhao Thalmann, Daniel Hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
description |
Vision-based hand pose estimation is important in human-computer interaction. While many recent works focus on full degree-of-freedom hand pose estimation, robust estimation of global hand pose remains a challenging problem. This paper presents a novel algorithm to optimize the leaf weights in a Hough forest to assist global hand pose estimation with a single depth camera. Different from traditional Hough forest, we propose to learn the vote weights stored at the leaf nodes of a forest in a principled way to minimize average pose prediction error, so that ambiguous votes are largely suppressed during prediction fusion. Experiments show that the proposed method largely improves pose estimation accuracy with optimized leaf weights on both synthesis and real datasets and performs favorably compared to state-of-the-art convolutional neural network-based methods. On real-world depth videos, the proposed method demonstrates improved robustness compared to several other recent hand tracking systems from both industry and academy. Moreover, we utilize the proposed method to build virtual/augmented reality applications to allow users to manipulate and examine virtual objects with bare hands. |
author2 |
School of Electrical and Electronic Engineering |
author_facet |
School of Electrical and Electronic Engineering Liang, Hui Yuan, Junsong Lee, Jun Ge, Liuhao Thalmann, Daniel |
format |
Article |
author |
Liang, Hui Yuan, Junsong Lee, Jun Ge, Liuhao Thalmann, Daniel |
author_sort |
Liang, Hui |
title |
Hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
title_short |
Hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
title_full |
Hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
title_fullStr |
Hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
title_full_unstemmed |
Hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
title_sort |
hough forest with optimized leaves for global hand pose estimation with arbitrary postures |
publishDate |
2020 |
url |
https://hdl.handle.net/10356/139907 |
_version_ |
1681058666189422592 |