الف نظامی
لائبریرین
Its an OCR that recognizes Urdu text optically from an offline image in any format and generates an Urdu document file easily editable in any word processor including MS Word.
It is generally based on the flexible Ligature Recognition Engine which can be used to develop any kind of customized text recognition software according to the specific needs of the client depending upon the font. It accepts input in the form of different image formats taken from an image capturing device like scanner. It extracts each ligature’s geometry from that input image and generates a polar map which is RTS invariant. This polar image is analyzed in the frequency domain. The Fourier coefficients are obtained from the so obtained frequencies. These coefficients are then used as features to recognize a specific ligature. Nastalique Font of Urdu is in progress and more than 1500 ligatures of different sizes have been tested using the technique with an accuracy of 92%.It is being developed as a Final Year Project and it is still in progress.
بحوالہIt is generally based on the flexible Ligature Recognition Engine which can be used to develop any kind of customized text recognition software according to the specific needs of the client depending upon the font. It accepts input in the form of different image formats taken from an image capturing device like scanner. It extracts each ligature’s geometry from that input image and generates a polar map which is RTS invariant. This polar image is analyzed in the frequency domain. The Fourier coefficients are obtained from the so obtained frequencies. These coefficients are then used as features to recognize a specific ligature. Nastalique Font of Urdu is in progress and more than 1500 ligatures of different sizes have been tested using the technique with an accuracy of 92%.It is being developed as a Final Year Project and it is still in progress.
