计算机与数字工程2019，Vol.47Issue(7)：1792-1796,5.DOI:10.3969/j.issn.1672-9722.2019.07.046
基于稀疏自编码器与梯度方向直方图的手势识别
Gesture Recognition scheme Based on Histogram of Oriented Gradients and Stacked Auto Encoder
摘要
Abstract
The traditional Histogram of Oriented Gradients(HOG)algorithm is sensitive to the contour edge information about or on the image during the extraction of image features,but the HOG feature lacks a description of the correlation between samples. Stacked Auto Encoder(SAE)can extract the characteristics of the sample of the feature extraction,but it requires a larger sample size. This paper combines HOG method and SAE method to use SVM classifier training classification gesture images under small da?ta sets,and experiments on the Jochen. Triesch data set show that this method can achieve better classification results.关键词
手势识别/稀疏自编码器/特征提取Key words
gesture recognition/sparse auto encoder/feature detection分类
信息技术与安全科学引用本文复制引用
缑新科,高庆东..基于稀疏自编码器与梯度方向直方图的手势识别[J].计算机与数字工程,2019,47(7):1792-1796,5.
