Feature Learning for Interaction Activity Recognition in RGBD Videos

TL;DR

This paper introduces a feature learning approach for human activity recognition in RGBD videos that outperforms existing methods by using a bag-of-visual-words model and SVM classification without relying on skeleton or tracking data.

Contribution

It presents a novel feature encoding method for activity recognition from RGBD videos that does not depend on domain-specific skeleton or tracking information.

Findings

Outperforms other techniques on RGBD activity datasets

Effective use of bag-of-visual-words for feature encoding

Achieves high recognition accuracy without skeleton data

Abstract

This paper proposes a human activity recognition method which is based on features learned from 3D video data without incorporating domain knowledge. The experiments on data collected by RGBD cameras produce results outperforming other techniques. Our feature encoding method follows the bag-of-visual-word model, then we use a SVM classifier to recognise the activities. We do not use skeleton or tracking information and the same technique is applied on color and depth data.