Pose-invariant kinematic features for action recognition

Recognition of actions from videos is a difficult task due to several factors like dynamic backgrounds, occlusion, pose-variations observed. To tackle the pose variation problem, we propose a simple method based on a novel set of pose-invariant kinematic features which are encoded in a human body ce...

Full description

Saved in:
Bibliographic Details
Main Authors: Ramanathan, Manoj, Yau, Wei-Yun, Teoh, Eam Khwang, Thalmann, Nadia Magnenat
Other Authors: School of Electrical and Electronic Engineering
Format: Conference or Workshop Item
Language:English
Published: 2020
Subjects:
Online Access:https://hdl.handle.net/10356/138068
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-138068
record_format dspace
spelling sg-ntu-dr.10356-1380682020-09-26T21:52:59Z Pose-invariant kinematic features for action recognition Ramanathan, Manoj Yau, Wei-Yun Teoh, Eam Khwang Thalmann, Nadia Magnenat School of Electrical and Electronic Engineering 2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC) Institute for Media Innovation (IMI) Engineering::Computer science and engineering Engineering::Electrical and electronic engineering Action Recognition Pose-invariance Recognition of actions from videos is a difficult task due to several factors like dynamic backgrounds, occlusion, pose-variations observed. To tackle the pose variation problem, we propose a simple method based on a novel set of pose-invariant kinematic features which are encoded in a human body centric space. The proposed framework begins with detection of neck point, which will serve as a origin of body centric space. We propose a deep learning based classifier to detect neck point based on the output of fully connected network layer. With the help of the detected neck, propagation mechanism is proposed to divide the foreground region into head, torso and leg grids. The motion observed in each of these body part grids are represented using a set of pose-invariant kinematic features. These features represent motion of foreground or body region with respect to the detected neck point's motion and encoded based on view in a human body centric space. Based on these features, poseinvariant action recognition can be achieved. Due to the body centric space is used, non-upright human posture actions can also be handled easily. To test its effectiveness in non-upright human postures in actions, a new dataset is introduced with 8 non-upright actions performed by 35 subjects in 3 different views. Experiments have been conducted on benchmark and newly proposed non-upright action dataset to identify limitations and get insights on the proposed framework. NRF (Natl Research Foundation, S’pore) ASTAR (Agency for Sci., Tech. and Research, S’pore) Accepted version 2020-04-23T04:31:11Z 2020-04-23T04:31:11Z 2018 Conference Paper Ramanathan, M., Yau, W.-Y., Teoh, E. K., & Thalmann, N. M. (2017). Pose-invariant kinematic features for action recognition. Proceedings of the 2017 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), 292-299. doi:10.1109/APSIPA.2017.8282038 9781538615430 https://hdl.handle.net/10356/138068 10.1109/APSIPA.2017.8282038 292 299 en © 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/APSIPA.2017.8282038 application/pdf
institution Nanyang Technological University
building NTU Library
country Singapore
collection DR-NTU
language English
topic Engineering::Computer science and engineering
Engineering::Electrical and electronic engineering
Action Recognition
Pose-invariance
spellingShingle Engineering::Computer science and engineering
Engineering::Electrical and electronic engineering
Action Recognition
Pose-invariance
Ramanathan, Manoj
Yau, Wei-Yun
Teoh, Eam Khwang
Thalmann, Nadia Magnenat
Pose-invariant kinematic features for action recognition
description Recognition of actions from videos is a difficult task due to several factors like dynamic backgrounds, occlusion, pose-variations observed. To tackle the pose variation problem, we propose a simple method based on a novel set of pose-invariant kinematic features which are encoded in a human body centric space. The proposed framework begins with detection of neck point, which will serve as a origin of body centric space. We propose a deep learning based classifier to detect neck point based on the output of fully connected network layer. With the help of the detected neck, propagation mechanism is proposed to divide the foreground region into head, torso and leg grids. The motion observed in each of these body part grids are represented using a set of pose-invariant kinematic features. These features represent motion of foreground or body region with respect to the detected neck point's motion and encoded based on view in a human body centric space. Based on these features, poseinvariant action recognition can be achieved. Due to the body centric space is used, non-upright human posture actions can also be handled easily. To test its effectiveness in non-upright human postures in actions, a new dataset is introduced with 8 non-upright actions performed by 35 subjects in 3 different views. Experiments have been conducted on benchmark and newly proposed non-upright action dataset to identify limitations and get insights on the proposed framework.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Ramanathan, Manoj
Yau, Wei-Yun
Teoh, Eam Khwang
Thalmann, Nadia Magnenat
format Conference or Workshop Item
author Ramanathan, Manoj
Yau, Wei-Yun
Teoh, Eam Khwang
Thalmann, Nadia Magnenat
author_sort Ramanathan, Manoj
title Pose-invariant kinematic features for action recognition
title_short Pose-invariant kinematic features for action recognition
title_full Pose-invariant kinematic features for action recognition
title_fullStr Pose-invariant kinematic features for action recognition
title_full_unstemmed Pose-invariant kinematic features for action recognition
title_sort pose-invariant kinematic features for action recognition
publishDate 2020
url https://hdl.handle.net/10356/138068
_version_ 1681057673783541760