OCR-based image compression

Apparatus for compressing a digitized image of a document, the apparatus comprising:

• OCR apparatus for performing optical character recognition (OCR) on the digitized image so as to assign alphanumeric character values to characters comprised in the image;
• a class identification unit for identifying, based at least in part, on the alphanumeric character values output by the OCR apparatus, a plurality of classes of the characters comprised in the image, the characters in each said class of characters having a common one of the alphanumeric character values assigned thereto by the OCR apparatus and each said class comprising at least one character;
• a pruning unit for pruning each class of characters, thereby producing:
• information describing the plurality of classes of characters;
• and
• a residual image;
• and
• a utilization unit for utilizing the information describing the plurality of class of characters and the residual image as a compressed digitized image in further processing,
• wherein the OCR apparatus is arranged to assign respective confidence levels to the characters with respect to the character value assigned thereto, and wherein the class identification unit is arranged to exclude from the classes the characters whose confidence levels are below a predefined level, and
• wherein the utilization unit is arranged to select from each of the classes one of the characters whose respective confidence level is highest to serve as a template for the class, wherein the digitized image is decompressed by using the template in place of each of the characters in the class.