Document data compression system which automatically segments documents and generates compressed smart documents therefrom

A system for automatic segmentation of a scanned document into separate document areas containing different types of image information which can vary in gray scale value comprising:

• means for scanning the document to generate a scanned image thereof;
• means for subdividing the scanned image into a matrix of blocks;
• means for analyzing the information contained in each block to classify said blocks as active or non-active, therein providing an activity classification of said blocks;
• means for generating macroblocks by grouping one or more adjacent said blocks in said matrix in accordance with their activity classification thereby segmenting said document;
• and
• means for analyzing the information contained in the blocks composing the macroblocks to classify each said macroblock as one of two image types in which when any of said macroblocks have at least one area within one or more blocks of the macroblock that has information which transitions outside a range from each of two gray scale values representative of the macroblock, the macroblock is a second of said two image types, and when said macroblock lacks at least one said area, said macroblock is a first of said two image types.

A method for producing a compressed document from a scanned document in a document processing device in which said scanned document has one or more different types of image information which can vary in gray scale value, said method comprising the steps of:

• subdividing the scanned document into a matrix of blocks;
• segmenting the scanned document into at least one group of one or more of said blocks which are adjacent to each other in said matrix in accordance with the image information of said scanned document to provide one or more macroblocks;
• locating one or more regions of one or more said blocks which are adjacent to each other in said matrix and are not part of said macroblocks;
• determining a level for each region corresponding to a numerical characteristic of the information in said region;
• analyzing the information in the blocks of said macroblocks to classify each macroblock as first or second types in which when any of said macroblocks have at least one area within one or more blocks of the macroblock that has information which transitions outside a range from each of prevalent two gray scale values of the macroblock, which are sufficiently different from each other, the macroblock is a second type macroblock, and when said macroblocks lacks at least one said area, the macroblock is a first type macroblock;
• producing a map representing the blocks in said matrix of said first type macroblocks;
• generating a binary image of the information in the blocks represented in said map;
• determining two levels for each said first type macroblock corresponding to the numerical characteristic of the information therein;
• determining position data specifying the locations of the macroblocks of said second type in said matrix;
• and
• generating said compressed document comprising said binary image, the information in said second type macroblocks, said position data, said map, said level of each region, and said levels of said first type macroblocks.

A method for automatic segmentation of a scanned document in a document processing device to separate document areas containing different types of image information comprising the steps of:

• scanning the document to generate a scanned image thereof;
• subdividing the scanned image into a matrix of blocks;
• analyzing the information contained in each block to classify said blocks as active or non-active, therein providing an activity classification of said blocks;
• generating macroblocks by grouping one or more adjacent said blocks in said matrix in accordance with their activity classification;
• and
• analyzing the information contained in the blocks composing the macroblocks to classify each said macroblock as one of two image types in which when any of said macroblocks have at least one area within one or more blocks of the macroblock that has information which transitions outside a range from each two gray scale values representative of the macroblock, the macroblock is a second of said two image types, and when said macroblock lacks at least one said area, said macroblock is a first of said two image types.

A system for producing a smart document from a primitive document which can have different types of image information comprising:

• means for analyzing the information contained in regions of said primitive document to classify said regions as active or non-active, therein providing an activity classification of said regions;
• means for generating segments by grouping one or more adjacent said regions responsive to their activity classification;
• means for analyzing the information contained in the regions composing the segments to classic each segment as one of binarizable or non-binarizable types by determining the presence of areas of image transition within the information of the segment by determining the absence or presence, respectively, of one or more areas of image transition within the information of the segment in accordance with said areas having information being outside one or more ranges of at least two different gray scale values characterizing the information of the segment;
• means for compressing each segment in accordance with their type;
• and
• means for outputting information about said compressed regions with sufficient format information about said primitive document to provide said smart document.