| 摘要 |
A method for automatically recognizing Arabic text includes building an Arabic dataset comprising Arabic text files written in different writing styles and actual meanings of the Arabic text corresponding to each of the Arabic text files (100), storing writing-style indices in association with the Arabic text files (105), digitizing a line of Arabic characters to form an array of pixels (321-323) (130), dividing the line of the Arabic characters into line images, (311-313) (120), forming a text feature vector from the line images (311-313) (140), training a Hidden Markov Model using the Arabic text files and ground truths in the Arabic dataset in accordance with the writing-style indices (160), and feeding the text feature vector into a Hidden Markov Model to recognize the line of Arabic characters (170). |
