Title: The research of human interaction recognition based on fusion features of key frame feature library

Authors: Haiyan Zhang; Shangbing Gao

Addresses: Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huai'an, China ' Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huai'an, China; Nanjing Key Laboratory of Trusted Cloud Computing and Big Data Analysis, Nanjing Xiaozhuang University, Nanjing, China

Abstract: Some issues such as computational complexity and low recognition accuracy still exist in human interaction recognition. In order to solve the problem, the paper has proposed innovative and effective method based on fixed features of key frame feature library. Firstly, GIST feature and HOG feature were extracted from the pre-processed videos. Secondly, training videos with different kinds of actions were clustered by the K-means algorithm respectively to get key frame feature of each action for constructing key frame feature library. And similarity measure was utilised to calculate the frequency of different key frames in every interactive video, and statistical histogram representation of interactive videos were obtained. Finally, the decision level fusion was achieved by using SVM classifier based on histogram intersection kernel to obtain impressive results on UT-interaction dataset. The correct recognition rate of the proposed algorithm is 85%, which indicates that the proposed algorithm is simple and effective than others.

Keywords: GIST feature; HOG feature; key frame feature library; histogram intersection kernel; UT-interaction dataset.

DOI: 10.1504/IJICT.2021.111918

International Journal of Information and Communication Technology, 2021 Vol.18 No.1, pp.57 - 69

Received: 26 Aug 2019
Accepted: 18 Oct 2019

Published online: 21 Dec 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article