Improvements in hidden Markov model based Arabic OCR

Statistical Markovian Data Modeling for Natural Language Processing

“…Reference [38] used HMM for recognizing Farsi handwritten words. Reference [39] describes recent advances in HMM based OCR for machine-printed Arabic documents. Reference [40] proposed a HMM based method for named entity recognition.…”

Open-vocabulary recognition of machine-printed Arabic text using hidden Markov models

“…Prasad et al [8] presented some improvements to the Arabic OCR system of the BBN Technologies. They presented the use of the parts of Arabic words (PAW) language models, which demonstrated better performance in terms of recognition rates, which exceeded the performance of word and character language models.…”

“…We employ a two-step approach in which the input text line image is associated with the closest known font in the first step and HMM-based text recognition is performed in the second step using the recognizer that was trained on the associated text font. This approach proved to be more effective than the commonly employed approach for recognizing text using a recognizer that was trained on text samples from multiple fonts [7], [8]. Our approach overcomes the common limitations of other techniques, such as the need for labeled samples of the text images in the font to be recognized and the assumption of data isogeny, i.e., the text lines to be recognized are obtained from only one font at a time [9].…”

A Case Study in Computer Understanding of Printed-Forms

“…Arabic and Farsi scripts are two well-known cursive scripts widely using. Reading these scripts requires an excessive segmentation process, say, word segmentation [23,24]. Technically speaking, a Farsi and/or Arabic word consists of connected characters (see figure 3).…”