Recognition of Odia Character in an Image by Dividing the Image into Four Quadrants

Abstract

This paper deals with optical character recognition of Odia characters written in a particular font family 'AkrutiOriAshok-99' with different font sizes 18, 20, 22, 24, 26, 28, 36, 48 and 72 in Bold style. The font 'AkrutiOriAshok-99' is a font from the typing software 'Akruti'. The basic idea behind the approach followed in this paper is the character decomposition into four quadrants and then extracting features from each quadrant. The image processing techniques like converting the image to gray, resizing of image and converting gray image to binary are used in this approach. The system explained in this paper has two major parts: DictionaryBuilding and FindingMatch. For DictionaryBuilding, dictionary of images which are created either by scanning a document or a document converted to image, both written in same font family in different sizes. The features are extracted from each image in any font size in the 'Dictionary' using Preprocessing, FindPath, GettingFeaturesLeft or GettingFeaturesRight, VisitSubQuad, RemainingSubQuad, WriteToExcel and CommonFeature modules. The part FindingMatch is responsible for finding a correct match in the dictionary for the input image. For this, FeatureExtraction and Recognition modules have been used. Longest Common Subsequence (LCS) has been used for finding the common feature in DictionaryBuilding as well as finding the correct match. A total of 1800 characters, 200 characters of each font size have been tested and 98.1% of correctness has been achieved.

An Approach to Recognise Odia Characters by Extracting Features from Fragmented Quadrants

“…Though the proposed system in this study follows some parts of the system described in (Kar and Pradhan, 2022), some other parts are newly added to the system. The system described in this study differs from the system described by (Kar and Pradhan, 2022) in the number of parts, division of the image of the Odia alphabet, direction which is followed to get a continuous trace of the black pixel, finding features of the images in 'Dictionary', input to the system and the correctness accuracy in the recognition of Odia alphabets; and the all the differences of the proposed system and the system described in the system (Kar and Pradhan, 2022) are tabulated in Table 1.…”

“…Table 1: Difference between the system described in this study and the system described in (Kar and Pradhan, 2022) Features of the system described in this study Features the system described in (Kar and Pradhan, 2022) 1.…”

“…For feature extraction from characters, different ways had been adopted in past works. The system in (Kar and Pradhan, 2022) created a dictionary of images of Odia alphabets in different font sizes and stored them in a directory Dictionary. The features were extracted from the images present in Dictionary by following some number of modules and the extracted features were written in an excel file, DictionaryFeatures.xlsx.…”

“…For recognition of characters based on the extracted features, the system in (Kar and Pradhan, 2022) searched the extracted features of the input image in the CommonFeature.xlsx excel file. The accuracy achieved in the recognition of the character was 98.1%.…”

“…The Build_Dictionary part has the following modules: Pre-Processing, Find_Direction, Get_Pattern_Left, Get_Pattern_Right, Sub_Quad_Pattern, Remaining_Part, Write_Excel, Find_Common_Feature, and Group_Vacant_Quads. The modules Get_Pattern_Left, Get_Pattern_Right, Sub_Quad_Pattern, Remaining_Part, and Write_Excel follow the steps of the modules GettingFeaturesLeft, Getting_Features_Right, Visit Sub Quad, Remaining Sub_Quad and Write_To_Excel of Feature Extraction in Dictionary_Building of the system in (Kar and Pradhan, 2022) respectively. The Find_Common_Feature module explained in this system, finds common features of each Odia alphabet by applying Longest Common Subsequence (LCS) on the features extracted from the Odia alphabet in all font sizes repeatedly.…”