An Algorithm for Japanese Character Recognition

Abstract

Abstract-In this paper we propose a geometry-topology based algorithm for Japanese Hiragana character recognition. This algorithm is based on center of gravity identification and is size, translation and rotation invariant. In addition, to the center of gravity, topology based landmarks like conjunction points masking the intersection of closed loops and multiple strokes, as well as end points have been used to compute centers of gravity of these points located in the individual quadrants of the circles enclosing the characters. After initial pre-processing steps like notarization, resizing, cropping, noise removal, synchronization, the total number of conjunction points as well as the total number of end points are computed and stored. The character is then encircled and divided into four quadrants. The center of gravity (cog) of the entire character as well as the cogs of each of the four quadrants are computed and the Euclidean distances of the conjunction and end points in each of the quadrants with the cogs are computed and stored. Values of these quantities both for target and template images are computed and a match is made with the character having the minimum Euclidean distance. Average accuracy obtained is 94.1 %.

Implementation of Zoning and K-Nearest Neighbor in Character Recognition of Wrésastra Script

“…A research about feature extraction and classification for X-Ray Medical Image [12] which proposed pertinent feature extraction algorithm for X-ray medical images and determined machine learning methods for automatic X-ray medical image classification. Previous research on Japanese Hiragana character recognition using Euclidean distance [13] obtained 94.1% average accuracy. Other research related to character recognition on Hindi Optical Character Recognition for Printed Documents using KNN [14] and a research about Javanese handwritten character classification [15].…”

Feature Extraction Techniques Based on Swarm Intelligence in OCR

“…They obtained the braces wise correspondences between word separators and unary assets in the word segmentation, because of the formulation as a binary quadratic algorithm and estimated the factors with the structured learning method [7]. Soumendu Das [8] explains Japanese Hiragana character recognition based on geometry topology. This algorithm explicit center of gravity identification, conversion and revolution invariant.…”

Japanese Historical Character Recognition using Deep Convolutional Neural Network (DCNN) with DropBlock Regularization

“…A geometry-topology based algorithm used to identify the Japanese hiragana characters and performed with an averaged accuracy rate of 94.1% [8]. In the state-of-art recognition, a Deep Convolutional Neural Network achieved an accuracy of 97.20% and 96.87% on CASIA-OLHWDBl.0 and CASIA-OLHWDBl.1 dataset to recognize Chinese handwritten character recognition [9].…”