Human Action Recognition Based on AdaBoost Algorithm for Feature Extraction

Author

Xiaofei Ji ; Lu Zhou ; Yibo Li

Author_Institution

Sch. of Autom., Shenyang Aerosp. Univ., Shenyang, China

fYear

2014

fDate

11-13 Sept. 2014

Firstpage

801

Lastpage

805

Abstract

A novel action recognition method based on AdaBoost algorithm is proposed in this paper. The method can select the most discriminative sample subset from a large amount of raw features of training data, so it can reduce the recognition computational complexity with high accuracy. The histogram of oriented gradient feature (HOG) descriptor is utilized to represent raw feature data. In order to select the most discriminative samples, Gadabouts algorithm is used to extract the raw feature data. The nearest neighbor classifier algorithm is utilized to test the proposed method on the UCF Sports database. Experiment results show that the method not only achieve the better recognition rate but also greatly improve the speed of recognition.

Keywords

feature extraction; feature selection; image classification; image motion analysis; image representation; learning (artificial intelligence); AdaBoost algorithm; Gadabouts algorithm; HOG descriptor; UCF Sports database; discriminative sample subset; feature data extraction; feature data representation; features selection; histogram of oriented gradient feature descriptor; human action recognition; nearest neighbor classifier algorithm; recognition computational complexity; Accuracy; Algorithm design and analysis; Classification algorithms; Educational institutions; Feature extraction; Histograms; Training; AdaBoost algorithm; Histogram of oriented gradient; Nearest neighbor classifier; UCF Sports database;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer and Information Technology (CIT), 2014 IEEE International Conference on

Conference_Location

Xi´an

Type

conf

DOI

10.1109/CIT.2014.87

Filename

6984755