Texture for script identification

• Published in: IEEE transactions on pattern analysis and machine intelligence Vol. 27; no. 11; pp. 1720 - 1732
• Main Authors: Busch, A., Boles, W.W., Sridharan, S.
• Format: Journal Article
• Language: English
• Published: Los Alamitos, CA IEEE 01.11.2005; IEEE Computer Society; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Applied sciences; Artificial Intelligence; Automatic Data Processing - methods; Character recognition; classification and association rules; clustering; Computer science; control theory; systems; document analysis; Documentation - methods; Exact sciences and technology; Handwriting; Image analysis; Image databases; Image Enhancement - methods; Image Interpretation, Computer-Assisted - methods; Index Terms- Script identification; Indexing; Information Storage and Retrieval - methods; Learning and adaptive systems; Models, Statistical; Numerical Analysis, Computer-Assisted; Optical character recognition; Optical character recognition software; Pattern Recognition, Automated - methods; Reading; Reproducibility of Results; Scripts; Sensitivity and Specificity; Signal Processing, Computer-Assisted; Sorting; Spatial databases; Subtraction Technique; Surface layer; Text analysis; Texture; Training data; Visual databases; wavelets and fractals; Cluster analysis; Data type; Fractals; Data mining; Precursor; Texture; Fractal; Classification; Database; Pattern analysis; Document analysis; Association rule; Image databank; classification and association rules; Statistical association; Text; Character recognition; Sorting; wavelets and fractals; Wavelet transformation; Optical character recognition; Script identification; clustering; Artificial intelligence; Indexing
• ISSN: 0162-8828; 1939-3539
• DOI: 10.1109/TPAMI.2005.227

The problem of determining the script and language of a document image has a number of important applications in the field of document analysis, such as indexing and sorting of large collections of such images, or as a precursor to optical character recognition (OCR). In this paper, we investigate the use of texture as a tool for determining the script of a document image, based on the observation that text has a distinct visual texture. An experimental evaluation of a number of commonly used texture features is conducted on a newly created script database, providing a qualitative measure of which features are most appropriate for this task. Strategies for improving classification results in situations with limited training data and multiple font types are also proposed.