Wipro Limited
Method and system for identifying bold text in a digital document

Abstract:

Disclosed herein is a method and device for identifying bold text in a digital document. The system receives image of digital document which comprises text. The system applies bounding box for each text in the image and scans predefined number of lines in each bounding box to identify width values of pixels. Thereafter, system identifies most occurring width value of pixels among the width values of pixels in each bounding box. The most occurring width value of pixels in each bounding box is identified as box width of corresponding bounding box. The system compares box width of each bounding box with threshold box width. If box width is greater than threshold box width, system identifies text of the bounding box whose box width exceeds threshold box width as bold text. The present disclosure efficiently identifies bold text in digital document based on width values of pixels with less computational power.