
For FullText PDF, please login, if you are a member of IEICE,
or go to Pay Per View on menu list, if you are a nonmember of IEICE.

Dimensionality Reduction for Histogram Features Based on Supervised Nonnegative Matrix Factorization
Mitsuru AMBAI Nugraha P. UTAMA Yuichi YOSHIDA
Publication
IEICE TRANSACTIONS on Information and Systems
Vol.E94D
No.10
pp.18701879 Publication Date: 2011/10/01
Online ISSN: 17451361 Print ISSN: 09168532 Type of Manuscript: Special Section PAPER (Special Section on InformationBased Induction Sciences and Machine Learning) Category: Keyword: dimensionality reduction, nonnegative matrix factorization, histogrambased features,
Full Text: PDF(928.7KB) >>Buy this Article
Summary:
Histogrambased image features such as HoG, SIFT and histogram of visual words are generally represented as highdimensional, nonnegative vectors. We propose a supervised method of reducing the dimensionality of histogrambased features by using nonnegative matrix factorization (NMF). We define a cost function for supervised NMF that consists of two terms. The first term is the generalized divergence term between an input matrix and a product of factorized matrices. The second term is the penalty term that reflects prior knowledge on a training set by assigning predefined constants to cannotlinks and mustlinks in pairs of training data. A multiplicative update rule for minimizing the newlydefined cost function is also proposed. We tested our method on a task of scene classification using histograms of visual words. The experimental results revealed that each of the lowdimensional basis vectors obtained from the proposed method only appeared in a single specific category in most cases. This interesting characteristic not only makes it easy to interpret the meaning of each basis but also improves the power of classification.

