eCite Digital Repository

Hierarchical topic modeling with Pose-transition feature for action recognition using 3D skeleton data


Huynh-The, T and Hua, C-H and Tu, NA and Hur, T and Bang, J and Kim, D and Amin, MB and Kang, BH and Seung, H and Shin, S-Y and Kim, E-S and Lee, S, Hierarchical topic modeling with Pose-transition feature for action recognition using 3D skeleton data, Information Sciences, 444 pp. 20-35. ISSN 0020-0255 (2018) [Professional, Refereed Article]

Pending copyright assessment - Request a copy

DOI: doi:10.1016/j.ins.2018.02.042


Despite impressive achievements in image processing and artificial intelligence in the past decade, understanding video-based action remains a challenge. However, the intensive development of 3D computer vision in recent years has brought more potential research opportunities in pose-based action detection and recognition. Thanks to the advantages of depth camera devices like the Microsoft Kinect sensor, we developed an effective approach to in-depth analysis of indoor actions using skeleton information, in which skeleton-based feature extraction and topic model-based learning are two major contributions. Geometric features, i.e. joint distance, joint angle, and joint-plane distance are calculated in the spatio-temporal dimension. These features are merged into two types, called pose and transition features, and then are provided to codebook construction to convert sparse features into visual words by k-means clustering. An efficient hierarchical model is developed to describe the full correlation of feature - poselet - action based on Pachinko Allocation Model. This model has the potential to uncover more hidden poselets, which have been recognized as the valuable information and help to differentiate pose-sharing actions. The experimental results on several well-known datasets, such as MSR Action 3D, MSR Daily Activity 3D, Florence 3D Action, UTKinect-Action 3D, and NTU RGB+D Action Recognition, demonstrate the high recognition accuracy of the proposed method. Our method outperforms state-of-the-art methods in the field in most dataset benchmarks.

Item Details

Item Type:Professional, Refereed Article
Keywords:3D action recognition, topic modeling, pose-transition feature, Pachinko Allocation Model, depth camera.
Research Division:Information and Computing Sciences
Research Group:Data management and data science
Research Field:Information retrieval and web search
Objective Division:Information and Communication Services
Objective Group:Information systems, technologies and services
Objective Field:Information systems, technologies and services not elsewhere classified
UTAS Author:Amin, MB (Dr Muhammad Bilal Amin)
UTAS Author:Kang, BH (Professor Byeong Kang)
ID Code:124914
Year Published:2018
Web of Science® Times Cited:28
Deposited By:Information and Communication Technology
Deposited On:2018-03-19
Last Modified:2021-03-23
Downloads:2 View Download Statistics

Repository Staff Only: item control page