An Optimization Framework for Image Contouring of Handwriting Recognition using Genetic Algorithm

Handwriting recognition poses a huge challenge in image processing due to the lack of defined character construction written by different people. This paper proposes a new technique addressing the research gap in the manual tuning enhancement of classification of handwritten Arabic numerals through introduction of image contouring in the pre-processing stage and its optimization using genetic algorithm (GA). Moreover, the study utilized image contouring techniques in extracting features of the image. Furthermore, the image contouring process was assisted by GA to optimize the tunings of the contours to maximize the features extracted. The experiment utilized the MNIST dataset for the training and testing. The accuracy per character varies due to the number of variations of a given character. Total accuracy of 99% was obtained during the testing.