A Nearest Opposite Contour Pixel Based Thinning Strategy for Character Images

A Nearest Opposite Contour Pixel Based Thinning Strategy for Character Images

Soumen Bag (Indian Institute of Technology (Indian School of Mines), Dhanbad, India)
Copyright: © 2017 |Pages: 18
DOI: 10.4018/978-1-5225-1025-3.ch006
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Thinning of character images is a big challenge. Removal of strokes or deformities in thinning is a difficult problem. In this paper, we have proposed a nearest opposite contour pixel based thinning strategy used for performing skeletonization of printed and handwritten character images. In this method, we have used shape characteristics of text to get skeleton of nearly same as the true character shape. This approach helps to preserve the local features and true shape of the character images. The proposed algorithm produces one pixel-width thin skeleton. As a by-product of our thinning approach, the skeleton also gets segmented into strokes in vector form. Hence further stroke segmentation is not required. Experiment is done on printed English and Bengali characters and we obtain less spurious branches comparing with other thinning methods without any post-processing.
Chapter Preview
Top

Introduction

Thinning is the process that reduces the amount of foreground shape information in images to facilitate efficient recognition and faster regeneration. One advantage of thinning is the Reduction of memory requirement for storing the essential structural information presented in a pattern. Moreover, it simplifies the data structure required in pattern analysis. Thinning of shape has a wide range of application in image processing, machine vision, and pattern recognition (Li & Basu, 1991; Sinha 1987). The ability to extract distinctive features from the pattern plays a key role in enhancing the effectiveness and efficiency in the task of recognizing patterns. In the domain of character images, strokes are the commonly extracted digital patterns used for recognition. Strokes can be easily extracted from a thinned character image. However, the process of thinning generally produces spurious strokes and shape deformations which later cause problems in the recognition task. Different thinning algorithms produce different degrees of distortion in character shape.

In the past several decades, many thinning algorithms have been developed (Lam et al., 1992; Vincze & K˝ov´ari, 2009). They are broadly classified into two groups: raster-scan based and medial-axis based. Raster-scan based methods are classified into two other categories:

  • Sequential, and

  • Parallel (Nagendraprasad et al., 1993).

Sequential algorithms consider one pixel at a time and visit all the pixels in the character by raster scanning or contour following (Arcelli, 1981; Arcelli & Baja, 1989; Beun, 1973; Rosenfeld, 1970; Wang & Zhang, 1989). Parallel thinning algorithms are based on iterative processing and they consider a pixel for removal based on the results of previous iteration. In 1966, Rutovitz (1966) proposed the first parallel thinning algorithm. Since then, many 2D parallel thinning algorithms have been proposed (Arcelli et al., 1975; Datta & Parui, 1994; Huang et al., 2003; Leung et al., 2000; Lu et al., 2010; Manzanera & Bernard, 2003; Pavlidis, 1982; Wang & Zhang, 1989; Zhang & Suen, 1984; Zhang & Wang, 1994, 1995; Zhu & Zhang, 2008).

A desirable property of thinning algorithms is to preserve the true shape of the given structure. But many of the raster-scan based character thinning methods cannot preserve the local properties or features of the character images properly (Couprie, 2005). As a result, they produce skeletons with some shape distortions. Medial-axis based methods generate a central or median line of the pattern directly in one pass without examining all the individual pixels (Martinez-Perez et al., 1987). However, the results suffer from local distortions especially near the junction points. Next, authors begin a brief description about few prominent thinning algorithms.

Top

Background

Thinning methods are classified into two types:

  • Raster-scan based, and

  • Medial-axis based.

Now, we will broadly discuss different types of thinning methods exist in literature.

Complete Chapter List

Search this Book:
Reset