Font classification using NMF

Abstract

In this paper, we propose a font classification method in scanned documents using non-negative matrix factorization (NMF). Using NMF, we automatically extract spatially local features enough to classify each font. The appropriateness of the features to classify a specific font is shown in the experimental results. The proposed method is expected to increase the performance of optical character recognition (OCR), document indexing and retrieval systems if such systems use a font classifier as a preprocessor.