A robust approach for text detection from natural scene images. Issue 9 (September 2015)

• Record Type: Journal Article
• Authors: Sun, Lei; Huo, Qiang; Jia, Wei; Chen, Kai
• Is Part Of: Pattern recognition. Volume 48:Issue 9(2015:Sep.)
• Page Start: 2906
• Page End: 2920
• Publication Date: 2015-09
• Subjects: Text detection -- Natural scene images -- Color-enhanced contrasting extremal region -- Neural networks; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2015.04.002
• Languages: English
• ISSNs: 0031-3203

Abstract

Abstract: This paper presents a robust text detection approach based on color-enhanced contrasting extremal region (CER) and neural networks. Given a color natural scene image, six component-trees are built from its grayscale image, hue and saturation channel images in a perception-based illumination invariant color space, and their inverted images, respectively. From each component-tree, color-enhanced CERs are extracted as character candidates. By using a "divide-and-conquer" strategy, each candidate image patch is labeled reliably by rules as one of five types, namely, Long, Thin, Fill, Square-large and Square-small, and classified as text or non-text by a corresponding neural network, which is trained by an ambiguity-free learning strategy. After pruning unambiguous non-text components, repeating components in each component-tree are pruned further. Remaining components are then grouped into candidate text-lines and verified by another set of neural networks. Finally, results from six component-trees are combined, and a post-processing step is used to recover lost characters. Our proposed method achieves superior performance on both ICDAR-2011 and ICDAR-2013 "Reading Text in Scene Images" test sets. Highlights: Several open problems in this topic are discussed and we present our solution. Color-enhanced CERs are effective to be candidate-text-connected-components. Neural networks work very well for the challenging text/non-text classification. The "ambiguity-free learning" strategy addresses the ambiguity problem properly. The "divide-and-conquer" strategy solves the size normalization problem well.