Exploiting zoning based on approximating splines in cursive script recognition

Abstract

Because of its complexity, handwriting recognition has to exploit many sources of information to be successful, e.g. the handwriting zones. Variability of zone-lines, however, requires a more flexible representation than traditional horizontal or linear methods. The proposed method therefore employs approximating cubic splines. Using entire lines of text rather than individual words is shown to improve the zoning accuracy, especially for short words. The new method represents an improvement over existing methods in terms of range of applicability, zone-line precision and zoning-classification accuracy. Application to several problems of handwriting recognition is demonstrated and evaluated.

Improving patch-based scene text script identification with ensembles of conjoined networks

“…Contrary to the case of printed document images, research in script identification on non traditional paper layouts is more scarce, and has been mainly dedicated to handwritten text [29,30,31,32,33], and video overlaidtext [12,34,35,36,13] until very recently. Gllavatta et al [12], in the first work dealing with video text script identification, proposed a method using the wavelet transform to detect edges in overlaid-text images.…”

A Comparison between the Performances of Several Hybrid Methods of Features Extraction for Isolated Printed Tifinagh Characters Recognition

“…More precisely, feature extraction methods can be divided into two principal categories: structural [7,15] and statistical [1,6] features The first category is based on local structure of numeral image while the second is interested to statistical information's localized in character image by way of example within this context there are the moments of images especially those invariants.…”

A Comparison between the Performances of Several Distances for Isolated Handwritten Arabic Numerals Recognition

“…Otherwise, handwritten character recognition is considered as a very challenging task given that the multitude of variability of writing styles from one person to another even just for a given person. On the other hand, several research papers has published for recognition of isolated handwritten Arabic or Latin character or numerals by using the zoning method [10][11][12][13][14][15] or the k nearest neighbors [1][2][3][4][5][6][7] Even so, this research is interested to isolated handwritten Arabic numerals recognition extracted from Mnist database [8,9]. Moreover, a set of three principal phases can be partitioned each OCR system.…”