New optimized text localization and character segmentation algorithms for Slab identification numbers

Abstract

This paper describes application-oriented text localization and character segmentation in images. Target texts in our application are often damaged by lots of various noise and, as a result, there are many unclear characters. Therefore, some special treatments are necessary to recognize these texts. In this paper, new optimized text localization and character segmentation algorithms are proposed. Experiments show that the proposed algorithms are fast and reliable.