An Evaluation of Local Action Descriptors for Human Action Classification in the Presence of Occlusion

2014 ◽  
pp. 56-67 ◽  
Author(s):  
Iveel Jargalsaikhan ◽  
Cem Direkoglu ◽  
Suzanne Little ◽  
Noel E. O’Connor
Keyword(s):  
Human Action ◽  
Local Action ◽  
Action Classification

scholarly journals Local Ternary Patterns from Three Orthogonal Planes for human action classification

2011 ◽  
Vol 38 (5) ◽  
pp. 5125-5128 ◽  
Author(s):  
Loris Nanni ◽  
Sheryl Brahnam ◽  
Alessandra Lumini
Keyword(s):  
Human Action ◽  
Action Classification ◽  
Local Ternary Patterns

An Action Classification Based on 2D2MSD

2013 ◽  
Vol 753-755 ◽  
pp. 3064-3067
Author(s):  
Ju Zhong ◽  
Ye Zi Sheng ◽  
Chun Li Lin ◽  
Nai Dong Cui
Keyword(s):  
Sample Size ◽  
Small Sample Size ◽  
Recognition Rate ◽  
Human Action ◽  
Small Sample ◽  
Experimental Results ◽  
Two Dimensional ◽  
Small Sample Size Problem ◽  
Action Classification ◽  
Average Recognition Rate

Double-direction two-dimensional Maximum Scatter Difference (2D2MSD) based on Maximum Scatter Difference (MSD) was proposed,which overcame the small sample size problem of LDA, and data were more concise. In the Weizmann human action database, experimental results showed the algorithm was fast, the average recognition rate reached 92% and the highest recognition rate reached 100%.

scholarly journals A Bayesian Dynamical Approach for Human Action Recognition

Sensors ◽  
2021 ◽  
Vol 21 (16) ◽  
pp. 5613
Author(s):  
Amirreza Farnoosh ◽  
Zhouping Wang ◽  
Shaotong Zhu ◽  
Sarah Ostadabbas
Keyword(s):  
Action Recognition ◽  
Large Scale ◽  
Temporal Dynamics ◽  
Human Action Recognition ◽  
Human Action ◽  
Superior Performance ◽  
Action Classification ◽  
Motion Data ◽  
Highly Correlated ◽  
Low Dimensional

We introduce a generative Bayesian switching dynamical model for action recognition in 3D skeletal data. Our model encodes highly correlated skeletal data into a few sets of low-dimensional switching temporal processes and from there decodes to the motion data and their associated action labels. We parameterize these temporal processes with regard to a switching deep autoregressive prior to accommodate both multimodal and higher-order nonlinear inter-dependencies. This results in a dynamical deep generative latent model that parses meaningful intrinsic states in skeletal dynamics and enables action recognition. These sequences of states provide visual and quantitative interpretations about motion primitives that gave rise to each action class, which have not been explored previously. In contrast to previous works, which often overlook temporal dynamics, our method explicitly model temporal transitions and is generative. Our experiments on two large-scale 3D skeletal datasets substantiate the superior performance of our model in comparison with the state-of-the-art methods. Specifically, our method achieved 6.3% higher action classification accuracy (by incorporating a dynamical generative framework), and 3.5% better predictive error (by employing a nonlinear second-order dynamical transition model) when compared with the best-performing competitors.

Sign in / Sign up

Export Citation Format

Share Document