Optical character recognition (OCR) systems are well-knownand very effectivein the area of the majorityof trendy language recognitions in present data. Not like other languages, the recognition of the Tamil language is highly difficult and thereforesignificantendeavorshave been put in state-of-the-art. Nevertheless, in order to recognize the Tamil characters in an accurate manner, the techniques are not so far developed. Hence, this paper presents a new Tamil Handwritten Character recognition model using 2 important procedures such as recognition as well as pre-processing. The conversion of RGB to grayscale is performed by the pre-processing stage, morphological operations image complementation, binarization with thresholding, as well as linearization. Subsequent to the linearization, the pre-processed image is fed to the recognition through an optimally configured K-Nearest Neighbour. Moreover, the Grey Wolf Optimization (GWO) algorithm is exploited to fine-tune the weights. The developed model performance is evaluated over the conventional techniques regarding various metrics.
Buy article PDF