Enhancement of ancient Marathi script using improved binarization method

Abstract

The digital image processing has major attention towards the automatic optical character recognition system. The image analysis of ancient images is mainly focused on printed and inscription images. In printed image analysis the text is printed by using machine and in the inscription images text is written by different persons. Presently most of the research highlighted challenges related to the inscription image analysis. In the printed document images font style, variation, size, and writing direction is standard, but in case of inscription font style, size, variation and writing direction vary by person to person. Image enhancement deals with the improvement of the quality of a digital image. The quality is achieved by highlighting useful details and removing the redundant details. In this paper, image enhancement based on the Binarization technique is discussed for the improvement of ancient inscription. The binary image is an essential part of document analysis. The proposed method is composed using a colour threshold and median filter. The proposed method is tested on different sample images. To evaluate the performance of the proposed method, the MSE, PSNR, F-Measure, NRM, Accuracy is calculated. These results are compared with manual results obtained by archeologist. It is also compared with the Nick method in the literature. The results of the proposed method are effective than manual and nick method. The enhancement of ancient Marathi scripts plays an important role in the identification, classification and the recognition of ancient Marathi characters.