A Review on Segmentation of Touching and Broken Characters for Handwritten Gurmukhi Script

Abstract

Character Segmentation of Handwritten Documents has been an interested area of research and its applicable environment becomes it a challenging research topic. The desire to edit the scanned document leads to develop the idea of optical character recognition. Segmentation plays very important role in optical character recognition system. The incorrect segmentation is just like a garbage in and garbage out. Segmentation of the broken character is quite difficult because vertical profile projection technique assumes the broken parts of the characters as individual characters. Existing methods focuses only upon the single touching characters. But our main focus is to design a robust method for the segmentation of broken and multiple touching characters. Existing systems focus only on the segmentation of fixed sized characters. But we develop the size independent algorithm which works on variable size characters. Thus, in the proposed method we develop the algorithm which works on the segmentation of broken, multiple touching characters of independent size including the three zones of the handwritten gurmukhi script. The main challenges like as variation in handwriting style etc. make the segmentation to difficult. General TermsOptical Character Recognition

Broken Character of Car Plate Recognition using Template Matching with Relocation Process

“…The process starts with pre-processing, character extraction by using morphological process and CCL, character identification and manual relocation of segmented broken character. Each of the character need to be identified to determine which one of the segmented character is the broken one using (3).…”

“…The example of the split broken character is shown in Figure 1. In Figure 1(a), the letter 'N' is broken at the middle of the letter, this case is in the category where the broken letter is possible to reconnect or reconstruct in order to form the original shape of 'N' by using a method suggested by using graph combinatorics method [1], Fuzzy Logical approach [2] and modified vertical projection profile [3]. Furthermore, by observing the letter, the actual behavior of letter is already known and easily to be guess because of the little broken part.…”

An Approach for Offline Handwritten Character Shape Reconstruction Using Active Contour and Morphological Techniques

“…Novak et al [17] have worked on the distortions present in the text printed on the metal ingot due to the uneven surface of the ingot for improved recognition by using fuzzy logic and human-inspired approach. Kaur and Bathla [18] have studied different techniques used for the segmentation of broken and touching characters written in Gurmukhi script. Yu and Yan [19] have implemented the dilation algorithm for the reconstruction of handwritten digits.…”