Image Based Hieroglyphic Character Recognition

摘要

Image Based Hieroglyphic Character Recognition was thought of to enable anyone interested in knowing the meaning of the hieroglyphs to use an algorithm to recognize the hieroglyphs to a well-known language. Since English is the most frequently used language in scientific work, therefore, the hieroglyphs will be translated to English language. The algorithm used is mainly about Optical Character Recognition (OCR) in the image processing field. The algorithm works as follows: An image that contains the hieroglyphs to be translated is taken as an input. Consequently, segmentation of the image will occur to cut every hieroglyph into a separate image, then, post-processing will be done to get only the region of interest in the image so that every image will be taken and compared to images in the data set to find the best match of the image using matching techniques. There were plenty of matching techniques tested until reaching Histogram of Oriented Gradients (HOG) that gave the best results in terms of accuracy. Then, the image will be translated to English language and displayed for the user in a text file. This paper addresses the contribution which is mainly controlling of the segmentation order for correct reading order by means of linkage to Gardiner's code and matching which is extremely essential to have correct results in recognition.