Stacked Denoising Autoencoder for feature representation learning in pose-based action recognition

Author

Budiman, A. ; Fanany, M.I. ; Basaruddin, C.

Author_Institution

Fac. of Comput. Sci., Univ. of Indonesia, Depok, Indonesia

fYear

2014

fDate

7-10 Oct. 2014

Firstpage

684

Lastpage

688

Abstract

In this paper, we studied Stacked Denoising Autoencoder(SDA) model for Human pose-based action recognition. We used public dataset Chalearn 2013 which contains Italian body language actions from 27 persons. We studied two model of SDA for pose clustering: 1) Traditional SDA with epoch and Neural Network supervised classifier and 2) Marginalized SDA which faster and ELM supervised classifier. We used supervised classifier by using initial cluster data from K-means. We deployed global tuning that updating the weight during iterative learning.

Keywords

feature extraction; image classification; image denoising; image motion analysis; learning (artificial intelligence); multilayer perceptrons; pattern clustering; pose estimation; ELM supervised classifier; Italian body language actions; SDA model; epoch classifier; feature representation learning; human pose-based action recognition; iterative learning; k-means clustering; marginalized SDA; multilayer perceptron; neural network supervised classifier; pose clustering; stacked denoising autoencoder; Accuracy; Joints; Neural networks; Noise reduction; Training; Tuning;

fLanguage

English

Publisher

ieee

Conference_Titel

Consumer Electronics (GCCE), 2014 IEEE 3rd Global Conference on

Conference_Location

Tokyo

Type

conf

DOI

10.1109/GCCE.2014.7031302

Filename

7031302