Artificial Intelligence


Human Activity Recognition Using Temporal Frame Decision Rule Extraction

Authors: Amol Patwardhan

Activities of humans and their recognition has many practical and real world applications such as safety, security, surveillance, humanoid assistive robotics and intelligent simulation systems. Numerous human action and emotion recognition systems included analysis of position and geometric features and gesture based co-ordinates to detect actions. There exits additional data and information in the movement and motion based features and temporal and time-sequential series of image and video frames which can be leveraged to detect and extract a certain actions, postures, gestures and expressions. This paper uses dynamic, temporal, time-scale dependent data to compare with decision rules and templates for activity recognition. The human shape boundaries and silhouette is extracted using geometric co-ordinate and centroid model across multiple frames. The extracted shape boundary is transformed to binary state using eigen space mapping and parameter dependent canonical transformation in 3D space dimension. The image blob data frames are down sampled using activity templates to a single candidate reference frame. This candidate frame was compared with the decision rule driven model to associate with an activity class label. The decision rule driven and activity templates method produced 64% recognition accuracy indicating that the method was feasible for recognizing human activities.

Comments: 4 Pages.

Download: PDF

Submission history

[v1] 2016-08-08 22:15:06

Unique-IP document downloads: 30 times

Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.

comments powered by Disqus