King Fahd University of Petroleum and Minerals

King Abdulaziz City for Science and Technology

Irfan Ahmad

Sabri A Mahmoud

Sabri Abdullah Mahmoud Mohammed

A text-recognition system and method include a trained classifier configured to identify a font of a frame of text line image. The system also includes an adaptive sliding window configured to divide the frame into a plurality of cells. A first cell is located around a prominent writing line of the text line image. Additional cells are located above the prominent writing line and below the prominent writing line, such that each of the additional cells above the prominent writing line has a same percentage of ink-pixels and each of the additional cells below the prominent writing line has a same percentage of ink-pixels. The system also includes a font-specific feature parameters database configured for extraction of text features from each of the cells. The system also includes one or more trained font-specific recognizers configured to recognize the extracted text features using an associated font-specific recognizer for the identified font.

Adaptive sliding windows for text recognition

A digital document skew detection and correction system that uses horizontal pixel projection analysis to detect the prominent writing line of Arabic text. The system removes both small and large components and divides the document into segments. The system uses the pixel projection results at varying angles to pinpoint the highest horizontal pixel concentration corresponding to the writing line of the Arabic text to determine the skew angle for each divided document segment. The skew angle variable is averaged between the documents segments before being reattached; a user notification application is activated if the differences between skewed angles exceed a threshold. The original document is rotated in the opposite direction of the estimated skew angle to correct the skew.

Technique for skew detection of printed arabic documents

The Arabic bank check analysis and zone extraction method for courtesy amount, legal amount, and the date fields uses available information and performs adaptive local processing using projections and connected component analysis. The method performs skew correction and connected component analysis on the check image and removes components smaller than average size before identifying the regions using projection profiles. This leads to a highly effective and robust technique. The method is based on prior knowledge of the check zone structure.

Growing community of inventors

Dhahran, Saudi Arabia

Irfan Ahmad