Title :
2DPCA-based method for place classification using range scan
Author :
Park, Soojin ; Park, Sung-Kee
Author_Institution :
Robot. Inst., Carnegie Mellon Univ., Pittsburgh, PA, USA
Abstract :
A method for place classification from a range scan is proposed, where the range scan is characterised with a 2D histogram of pairwise angular distance. The proposed method uses two-dimensional principal component analysis (2DPCA), a feature extraction and data representation technique, to train a set of classifiers with 2D histograms of range scans.
Keywords :
feature extraction; image classification; learning (artificial intelligence); principal component analysis; 2D histogram; 2DPCA; 2DPCA-based method; data representation technique; feature extraction; pairwise angular distance; place classification; range scan; supervised learning; two-dimensional principal component analysis;
Journal_Title :
Electronics Letters
DOI :
10.1049/el.2011.2907
