Document image processing method and system having function of determining body text region reading order

Claim 2

A document image processing method comprising machine implemented steps of: a) extracting text regions from an input document image; b) classifying said text regions into: (b1) in-order reading regions of text which are to be successively read in a predetermined order and (b2) different-attribute regions of text; c) detecting a construction of said in-order reading regions but not of said different-attribute regions; and d) determining the reading order, in which said in-order reading regions are to be read, using said construction.

Claim 3

The document image processing method according to claim 1, wherein: said detecting step c) detects said construction in a manner that is the same whether said input document image comprises a vertically typeset document or a horizontally typeset document.

Claim 4

The document image processing method according to claim 1, wherein said detecting step c) further includes a tree graph formation step c-1) forming a tree graph representing said construction including nodes respectively representing said in-order reading regions.

Claim 5

The document image processing method according to claim 1, wherein: said classifying step b) classifies said text regions into in-order reading regions enclosed by a box frame which is included in said input document image and in-order reading regions enclosed by no box frame; and said determining step determines the reading order in which said in-order reading regions enclosed by a box frame are to be read and further determines the reading order in which said in-order reading regions enclosed by no box frame are to be read.

Claim 6

The document image processing method according to claim 4, wherein: said box frame is formed of four sides consisting of a top side, a bottom side, a right side and a left side.

Claim 7

The document image processing method according to claim 1, further comprising: e) a checking step checking whether said reading order is correct or incorrect; and f) a re-determining step again determining the reading order if a result of incorrect is obtained in said checking step e).

Claim 8

The document image processing method according to claim 6, wherein: said checking step e) includes checking the reading order, in which the in-order reading regions are to be read, for the in-order reading regions enclosed in a box frame which is included in said input document image; and said re-determining step f) includes again determining the reading order for said in-order reading regions enclosed in said box frame if the re-determining operation is needed.

Claim 9

The document image processing method according to claim 7, wherein: said box frame is formed of four sides consisting of a top side, a bottom side, a right side and a left side.

Claim 10

The document image processing method according to claim 1, further comprising g) a character reading step reading the characters constituting said in-order reading regions in accordance with said reading order.

Claim 11

The document image processing method according to claim 1, wherein said detecting step c) includes: using imaginary ruled lines, which may be generated depending on contents of an original image and then added to said original image so as to control the detection operation.

Claim 12

The document image processing method according to claim 10, wherein said imaginary ruled lines are generated for blank regions having a predetermined shape/predetermined dimensions and/or for figure/table regions.

Claim 13

The document image processing method according to claim 1, wherein: said detecting step c) includes using scopes of influence, assigned to appropriate regions, for searching subsequent regions; and each of said scopes of influence assigned to a respective one of said appropriate regions represents a scope on which said respective one of said appropriate regions has influence.

Claim 14

The document image processing method according to claim 1, wherein: said different-attribute regions are prevented from affecting the determination of said determining step d).

Claim 15

The document image processing method according to claim 1, wherein said different-attribute regions include: a series of characters which are provided as footers and/or headers in said input document image.

Claim 16

A document image processing method comprising: a) extracting text regions from an input document image; b) classifying said text regions into (b1) in-order reading regions which are to be successively read in a predetermined order and into (b2) different-attribute regions; c) detecting a construction of said in-order reading regions; d) determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking whether said reading order is correct or incorrect; and f) re-determining the reading order if a result of incorrect is obtained in said checking step e); wherein said checking step e) includes: e1) providing reference points to the respective in-order reading regions; e2) connecting said reference points in accordance with a relevant reading order; and e3) determining said reading order to be incorrect if lines formed as the result of the connection intersect.

Claim 17

A document image processing method comprising: a) extracting text regions from an input document image; b) classifying said text regions into (b1) in-order reading regions which are to be successively read in a predetermined order and into (b2) different-attribute regions; c) detecting a construction of said in-order reading regions; d) determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking whether said reading order is correct or incorrect; and f) re-determining the reading order if a result of incorrect is obtained in said checking step e); wherein said checking step e) includes: e1) providing reference points to the respective in-order reading regions; e2) connecting said reference points in accordance with a relevant reading order; and e3) determining said reading order to be incorrect if lines formed as the result of the connection intersect.

Claim 18

A document image processing method comprising: a) extracting text regions from an input document image; b) classifying said text regions into (b1) in-order reading regions which are to be successively read in a predetermined order and into (b2) different-attribute regions; c) detecting a construction of said in-order reading regions; d) determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking whether said reading order is correct or incorrect; and f) re-determining the reading order if a result of incorrect is obtained in said checking step e); wherein said checking step e) includes: e1) providing reference points to the respective in-order reading regions; e2) connecting said reference points in accordance with a relevant reading order; and e3) determining said reading order to be incorrect if a number of intersections of the lines formed as a result of the connection exceeds a predetermined value.

Claim 19

A document image processing method comprising: a) extracting text regions from an input document image; b) classifying said text regions into (b1) in-order reading regions which are to be successively read in a predetermined order and into (b2) different-attribute regions; c) detecting a construction of said in-order reading regions; d) determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking whether said reading order is correct or incorrect; and f) re-determining the reading order if a result of incorrect is obtained in said checking step e); wherein said checking step e) includes: e1) providing reference points to the respective in-order reading regions; e2) connecting said reference points in accordance with a relevant reading order; and e3) determining said reading order to be incorrect if a number of intersections of the lines formed as a result of the connection exceeds a predetermined value.

Claim 20

A document image processing system comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into: (b1) in-order reading regions of text which are to be read in a predetermined order and (b2) different-attribute regions of text; c) detecting means for detecting a construction of said in-order reading regions but not of said different-attribute regions; and d) determining means for, without human intervention, determining the reading order, in which said in-order reading regions are to be read, using said construction.

Claim 21

A document image processing system comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into: (b1) in-order reading regions of text which are to be read in a predetermined order and (b2) different-attribute regions of text; c) detecting means for detecting a construction of said in-order reading regions but not of said different-attribute regions; and d) determining means for, without human intervention, determining the reading order, in which said in-order reading regions are to be read, using said construction.

Claim 22

The document image processing system according to claim 17, wherein: said detecting means detects said construction in a manner that is the same whether said input document image comprises a vertically typeset document or a horizontally typeset document.

Claim 23

The document image processing system according to claim 17, wherein said detecting means further includes a tree graph formation means for forming a tree graph representing said construction including nodes respectively representing said in-order reading regions.

Claim 24

The document image processing system according to claim 17, wherein: said classifying means classifies said text regions into in-order reading regions enclosed by a box frame which is included in said input document image and in-order reading regions enclosed by no box frame; and said determining means determines the reading order of said in-order reading regions enclosed by a box frame and further determines the reading order of said in-order reading regions enclosed by no box frame.

Claim 25

The document image processing system according to claim 20, wherein: said box frame is formed of four sides consisting of a top side, a bottom side, a right side and a left side.

Claim 26

The document image processing system according to claim 17, further comprising: checking means for checking whether said reading order is correct or incorrect; and re-determining means for again determining the reading order of said in-order reading regions using another predetermined procedure if a result of incorrect is obtained by said checking means.

Claim 27

The document image processing system according to claim 22, wherein: said checking means includes means for checking the reading order of the in-order reading regions for the in-order reading regions enclosed in a box frame which is included in said input document image; and said re-determining means again determines the reading order for said in-order reading regions enclosed in said box frame if the re-determining operation is needed to be performed.

Claim 28

The document image processing system according to claim 23, wherein: said box frame is formed of four sides consisting of a top side, a bottom side, a right side and a left side.

Claim 29

The document image processing system according to claim 17, further comprising character reading means for reading the characters constituting said in-order reading regions in accordance with said reading order.

Claim 30

The document image processing system according to claim 17, wherein said detecting means includes: means for using imaginary ruled lines, which may be generated depending on contents of an original image and then added to said original image so as to control the detection operation.

Claim 31

The document image processing method according to claim 26, wherein said imaginary ruled lines are generated for blank regions having a predetermined shape/predetermined dimensions and/or for figure/table regions.

Claim 32

The document image processing system according to claim 17, wherein: said detecting means includes means for using scopes of influence, assigned to appropriate regions, for searching subsequent regions; and each of said scopes of influence assigned to a respective one of said appropriate regions represents a scope on which said respective one of said appropriate regions has influence.

Claim 33

The document image processing system according to claim 17, wherein: said different-attribute regions are prevented from affecting the determination by said determining means.

Claim 34

The document image processing system according to claim 17, wherein said different-attribute regions include: a series of characters which are provided as footers and/or headers in said input document image.

Claim 35

A document image processing system comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into (b1) in-order reading regions which are to be read in a predetermined order and into (b2) different-attribute regions; c) detecting means for detecting a construction of said in-order reading regions; d) determining means for determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking means for checking whether said reading order is correct or incorrect; and f) re-determining means for again determining the reading order of said in-order reading regions using another predetermined procedure if a result of incorrect is obtained by said checking means; wherein said checking means includes: e1) means for providing reference points to the respective in-order reading regions; e2) means for connecting said reference points in accordance with a relevant reading order; and e3) means for determining said reading order to be incorrect if lines formed as the result of the connection intersect.

Claim 36

A document image processing system comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into (b1) in-order reading regions which are to be read in a predetermined order and into (b2) different-attribute regions; c) detecting means for detecting a construction of said in-order reading regions; d) determining means for determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking means for checking whether said reading order is correct or incorrect; and f) re-determining means for again determining the reading order of said in-order reading regions using another predetermined procedure if a result of incorrect is obtained by said checking means; wherein said checking means includes: e1) means for providing reference points to the respective in-order reading regions; e2) means for connecting said reference points in accordance with a relevant reading order; and e3) means for determining said reading order to be incorrect if lines formed as the result of the connection intersect.

Claim 37

A document image processing system comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into (b1) in-order reading regions which are to be read in a predetermined order and into (b2) different-attribute regions; c) detecting means for detecting a construction of said in-order reading regions; d) determining means for determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking means for checking whether said reading order is correct or incorrect; and f) re-determining means for again determining the reading order of said in-order reading regions using another predetermined procedure if a result of incorrect is obtained by said checking means; wherein said checking means includes: e1) means for providing reference points to the respective in-order reading regions; e2) means for connecting said reference points in accordance with a relevant reading order; and e3) means for determining said reading order to be incorrect if intersections of lines formed as the result of the connection exceeds a predetermined threshold value.

Claim 38

A document image processing system comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into (b1) in-order reading regions which are to be read in a predetermined order and into (b2) different-attribute regions; c) detecting means for detecting a construction of said in-order reading regions; d) determining means for determining the reading order, in which said in-order reading regions are to be read, using said construction; e) checking means for checking whether said reading order is correct or incorrect; and f) re-determining means for again determining the reading order of said in-order reading regions using another predetermined procedure if a result of incorrect is obtained by said checking means; wherein said checking means includes: e1) means for providing reference points to the respective in-order reading regions; e2) means for connecting said reference points in accordance with a relevant reading order; and e3) means for determining said reading order to be incorrect if intersections of lines formed as the result of the connection exceeds a predetermined threshold value.

Claim 39

A document image processing method, comprising machine implemented steps of: a) extracting text regions from an input document image; b) classifying said text regions into: (b1) body regions of text which are to be successively read in a predetermined order and (b2) different-attribute regions of text; c) detecting a construction of said body regions but not of said different-attribute regions; and d) determining the reading order, in which said body regions are to be read, using said construction.

Claim 40

A document image processing method, comprising machine implemented steps of: a) extracting text regions from an input document image; b) classifying said text regions into: (b1) body regions of text which are to be successively read in a predetermined order and (b2) different-attribute regions of text; c) detecting a construction of said body regions but not of said different-attribute regions; and d) determining the reading order, in which said body regions are to be read, using said construction.

Claim 41

A document image processing system, comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into: (b1) body regions of text which are to be successively read in a predetermined order and (b2) different-attribute regions of text; c) detecting means for detecting a construction of said body regions but not of said different-attribute regions; and d) determining means for, without human intervention, determining the reading order, in which said body regions are to be read, using said construction.

Claim 42

A document image processing system, comprising: a) extracting means for extracting text regions from an input document image; b) classifying means for classifying said text regions into: (b1) body regions of text which are to be successively read in a predetermined order and (b2) different-attribute regions of text; c) detecting means for detecting a construction of said body regions but not of said different-attribute regions; and d) determining means for, without human intervention, determining the reading order, in which said body regions are to be read, using said construction.