Segmentation of text and graphics

Claim 2

In a digital processing system, a method of identifying a line graphics region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structure element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said step of identifying comprising the steps of: i) forming a mask, the mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 3

The method as recited in claim 1 wherein the step of eliminating vertical lines from said first intermediate image further comprises the steps of: a) closing said image containing text and line graphics with a structuring element having at least two adjacent horizontal ON pixels; and b) opening said image with a structuring element having at least two adjacent horizontal ON pixels to produce said first intermediate image.

Claim 4

The method as recited in claim 2 wherein the structuring element used in said step of opening uses a structuring element having at least three adjacent horizontal ON pixels.

Claim 5

The method as recited in claim 3 wherein said closing step uses a structuring element with at least three adjacent horizontal ON pixels and said step of opening uses a structuring element having at least four adjacent horizontal ON pixels.

Claim 6

The method as recited on claim 2 preceded by the step of at least one reduction of said image containing text and line graphics.

Claim 7

The method as recited in claim 2 preceded by the step of at least two reductions of said image containing text and line graphics.

Claim 8

The method as recited in claim 1 wherein the step of turning on additional pixels is a step of turning pixels ON within a bounding box.

Claim 9

The method as recited in claim 7 wherein the step of tuning on additional pixels is a step of iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change.

Claim 10

The method as recited in claim 7 wherein the step of turning on additional pixels is preceded by an additional reduction step.

Claim 11

The method as recited in claim 7 wherein the step of turning on additional pixels is preceded by a closing step, said closing step using a structuring element (SE) having at least three adjacent vertical ON pixels.

Claim 12

The method as recited in claim 7 wherein the step of turning on additional pixels is followed by an opening step, said open using at least a 2×2 structuring element (SE) of solid ON pixels.

Claim 13

The method as recited in claim 7 followed by the step of expanding said mask by a factor of at least two.

Claim 14

The method as recited in claim 7 followed by the step of dilating said mask.

Claim 15

The method as recited in claim 7 followed by the step of: a) expanding said mask by a factor of at least two then; b) dilating said mask then; c) again expanding said mask by a factor of at least two.

Claim 16

In a digital processing system, a method of identifying a line graphics region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said step of identifying comprising the steps of: i) forming a mask, the mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of the horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 17

In a digital processing system, a method of identifying a line graphics region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said step of identifying comprising the steps of: i) forming a mask, the mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of the horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 18

The method as recited in claim 15 wherein the step of substantially eliminating horizontal lines from said image comprising the step of opening said image containing text and line graphics with a structuring element having at least two adjacent vertical ON pixels.

Claim 19

The method as recited in claim 15 wherein the step of substantially eliminating horizontal lines from said image comprises the steps of: a) closing said image containing text and line graphics, said step of closing using a structuring element having at least two adjacent vertical ON pixels; and b) opening said image, said step of opening using a structuring element having at least two adjacent vertical ON pixels to produce said first intermediate image.

Claim 20

The method as recited in claim 16 preceded by the steps of: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 21

The method as recited in claim 17 preceded by the steps of: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 22

A digital computer system programmed to identify a line graphics region in an image containing at least test and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 23

A digital computer system programmed to identify a line graphics region in an image containing at least test and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 24

The computer system as recited in claim 20 programmed to eliminate vertical lines from said first intermediate image by: a) closing said image containing text and line graphics with a structuring element having at least two adjacent horizontal ON pixels; and b) opening said image with a structuring element having at least two adjacent horizontal ON pixels to produce said first intermediate image.

Claim 25

The computer system as recited in claim 21 wherein the structuring element used to open has at least three adjacent horizontal ON pixels.

Claim 26

The computer system as recited in claim 21 wherein the structuring element used to close has at least three adjacent horizontal ON pixels and the structuring element used to open has at least one more adjacent horizontal ON pixel than said structuring element used to close.

Claim 27

The computer system as recited in claim 21 further programmed to perform at least one reduction of said image containing text and line graphics.

Claim 28

The computer system as recited in claim 21 further programmed to perform at least two reductions of said image containing text and line graphics.

Claim 29

The computer system as recited in claim 20 programmed to turn on additional pixels by turning pixels on up to a bounding box.

Claim 30

The computer system as recited in claim 26 programmed to turn on additional pixels by iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change.

Claim 31

The computer system as recited in claim 26 further programmed to perform an additional reduction before turning on additional pixels.

Claim 32

The computer system as recited in claim 26 further programmed to perform a closing step, said closing step using a structuring element (SE) having at least three adjacent vertical ON pixels.

Claim 33

The computer system as recited in claim 26 further programmed to perform an opening step, said opening using at least a 2×2 structuring element (SE) of solid ON pixels.

Claim 34

The computer system as recited in claim 26 further programmed to expand said mask by a factor of at least two.

Claim 35

The computer system as recited in claim 26 further programmed to dilate said mask.

Claim 36

The computer system as recited in claim 26 further programmed to: a) expand said mask then by a factor of at least two; b) dilate said mask then; c) again expand said mask by a factor of at least two.

Claim 37

A digital computer system programmed to identify a line graphics region in an image containing at least text and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 38

A digital computer system programmed to identify a line graphics region in an image containing at least text and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 39

The computer system as recited in claim 34 programmed to substantially eliminate horizontal lines from said image by opening said image containing text and line graphics with a structuring element having at least two adjacent vertical ON pixels.

Claim 40

The computer system as recited in claim 34 programmed to substantially eliminate horizontal lines from said image by: a) closing said image containing text and line graphics, said step of closing using a structuring element having at least two adjacent vertical ON pixels; and b) opening said image, said step of opening using a structuring element having at least two adjacent vertical ON pixels to produce said first intermediate image.

Claim 41

The computer system as recited in claim 35 further programmed to process the image by: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 42

The computer system as recited in claim 36 further programmed to process the image by: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 43

In a digital processing system, a method of identifying a line graphics region in an image containing at least text and line graphics, comprising the steps of converting OFF pixels adjacent text pixels to ON pixels, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels, and identifying at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said steps of converting and identifying further comprising the steps of: a) forming a mask, said mask comprising ON pixels in a region of text pixels, the step of forming a mask further comprising the steps of: i) substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image; and ii) turning on additional pixels adjacent to remaining ON pixels in said first intermediate image wherein the step of turning on additional pixels is a step of turning pixels on within a bounding box, by iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change, said step of iteratively eroding and dilating comprising the steps of: eroding said first intermediate image with a structuring element (SE) having ON pixels along a first diagonal to produce a second intermediate image; dilating said second intermediate image with a SE having ON pixels along a second diagonal to produce a third intermediate image; ORing said third intermediate image with said first intermediate image to produce a forth intermediate image; eroding said fourth intermediate image with said second SE to produce a fifth intermediate image; dilating said fifth intermediate image with said first SE to produce a sixth intermediate image; ORing said sixth intermediate image with said fourth intermediate image to produce a seventh intermediate image; and returning to said eroding step using said seventh intermediate image as said first intermediate image until said first intermediate image and said seventh intermediate image do not substantially differ; and b) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 44

In a digital processing system, a method of identifying a line graphics region in an image containing at least text and line graphics, comprising the steps of converting OFF pixels adjacent text pixels to ON pixels, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels, and identifying at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said steps of converting and identifying further comprising the steps of: a) forming a mask, said mask comprising ON pixels in a region of text pixels, the step of forming a mask further comprising the steps of: i) substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image; and ii) turning on additional pixels adjacent to remaining ON pixels in said first intermediate image wherein the step of turning on additional pixels is a step of turning pixels on within a bounding box, by iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change, said step of iteratively eroding and dilating comprising the steps of: eroding said first intermediate image with a structuring element (SE) having ON pixels along a first diagonal to produce a second intermediate image; dilating said second intermediate image with a SE having ON pixels along a second diagonal to produce a third intermediate image; ORing said third intermediate image with said first intermediate image to produce a forth intermediate image; eroding said fourth intermediate image with said second SE to produce a fifth intermediate image; dilating said fifth intermediate image with said first SE to produce a sixth intermediate image; ORing said sixth intermediate image with said fourth intermediate image to produce a seventh intermediate image; and returning to said eroding step using said seventh intermediate image as said first intermediate image until said first intermediate image and said seventh intermediate image do not substantially differ; and b) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 45

In a digital processing system, a method of identifying a text region in an image containing at least text and line graphics, comprising the steps of converting OFF pixels adjacent text pixels to ON pixels, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels, and identifying at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, the steps of converting and identifying further comprising the steps of: a) forming a mask, said mask comprising ON pixels in a region of text pixels, the step of forming a mask further comprising the steps of: i) substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image; and ii) turning on additional pixels adjacent to remaining ON pixels in said first intermediate image, wherein the step of turning on additional pixels is a step of turning pixels on within a bounding box, by iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change, said step of iteratively eroding and dilating comprising the steps of: eroding said first intermediate image with a structuring element (SE) having ON pixels along a first diagonal to produce a second intermediate image; dilating said second intermediate image with an SE having ON pixels along a second diagonal to produce a third intermediate image; ORing said third intermediate image with said first intermediate image to produce a fourth intermediate image; eroding said fourth intermediate image with said second SE to produce a fifth intermediate image; dilating said fifth intermediate image with said first SE to produce a sixth intermediate image; ORing said sixth intermediate image with said fourth intermediate image to produce a seventh intermediate image; and returning to said eroding step using said seventh intermediate image as said first intermediate image until said first intermediate image and said seventh intermediate image do not substantially differ; and b) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 46

In a digital processing system, a method of identifying a text region in an image containing at least text and line graphics, comprising the steps of converting OFF pixels adjacent text pixels to ON pixels, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels, and identifying at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, the steps of converting and identifying further comprising the steps of: a) forming a mask, said mask comprising ON pixels in a region of text pixels, the step of forming a mask further comprising the steps of: i) substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image; and ii) turning on additional pixels adjacent to remaining ON pixels in said first intermediate image, wherein the step of turning on additional pixels is a step of turning pixels on within a bounding box, by iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change, said step of iteratively eroding and dilating comprising the steps of: eroding said first intermediate image with a structuring element (SE) having ON pixels along a first diagonal to produce a second intermediate image; dilating said second intermediate image with an SE having ON pixels along a second diagonal to produce a third intermediate image; ORing said third intermediate image with said first intermediate image to produce a fourth intermediate image; eroding said fourth intermediate image with said second SE to produce a fifth intermediate image; dilating said fifth intermediate image with said first SE to produce a sixth intermediate image; ORing said sixth intermediate image with said fourth intermediate image to produce a seventh intermediate image; and returning to said eroding step using said seventh intermediate image as said first intermediate image until said first intermediate image and said seventh intermediate image do not substantially differ; and b) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 47

A computer programmed to perform the method of claims 39 or 40.

Claim 48

In a digital processing system, a method of identifying a text region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text then said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, the step of identifying further comprising the steps of: i) forming a mask, said mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 49

In a digital processing system, a method of identifying a text region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text then said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, the step of identifying further comprising the steps of: i) forming a mask, said mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 50

The method as recited in claim 42 wherein the step of eliminating vertical lines from said first intermediate image further comprising the steps of: a) closing said image containing text and line graphics with a structuring element having at least two adjacent horizontal ON pixels; and b) opening said image with a structuring element having at least two adjacent horizontal ON pixels to produce said first intermediate image.

Claim 51

The method as recited in claim 43 wherein the structuring element used in said step of opening uses a structuring element having at least three adjacent horizontal ON pixels.

Claim 52

The method as recited in claim 44 wherein said closing step uses a structuring element with at least three adjacent horizontal ON pixels and said step of opening uses a structuring element having at least four adjacent horizontal ON pixels.

Claim 53

The method as recited in claim 43 preceded by the step of at least one reduction of said image containing text and line graphics.

Claim 54

The method as recited in claim 43 preceded by the step of at least two reductions of said image containing text and line graphics.

Claim 55

The method as recited in claim 42 wherein the step of turning on additional pixels is a step of turning pixels ON within a bounding box.

Claim 56

The method as recited in claim 48 wherein the step of turning on additional pixels is a step of iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change.

Claim 57

The method as recited in claim 48 wherein the step of turning on additional pixels is preceded by an additional reduction step.

Claim 58

The method as recited in claim 48 wherein the step of turning on additional pixels is preceded by a closing step, said closing step using a structuring element (SE) having at least three adjacent vertical ON pixels.

Claim 59

The method as recited in claim 48 wherein the step of turning on additional pixels is followed by an opening step, said open using at least a 2×2 structuring element (SE) of solid ON pixels.

Claim 60

The method as recited in claim 48 followed by the step of expanding said mask by a factor of at least two.

Claim 61

The method as recited in claim 48 followed by the step of dilating said mask.

Claim 62

The method as recited in claim 48 followed by the steps of: a( expanding said mask by a factor of at least two then; b) dilating said mask then; c) again expanding said mask b a factor of at least two.

Claim 63

In a digital processing system, a method of identifying a text region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, the step of identifying further comprising the steps of: i) forming a mask, said mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of the horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 64

In a digital processing system, a method of identifying a text region in an image containing at least text and line graphics, comprising the steps of: a) converting OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identifying at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, the step of identifying further comprising the steps of: i) forming a mask, said mask comprising ON pixels in a region of text pixels, said step of forming a mask comprising the steps of substantially eliminating at least a portion of the horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 65

The method as recited in claim 56 wherein the step of substantially eliminating horizontal lines from said image comprises the step of opening said image containing text and line graphics with a structuring element having at least two adjacent vertical ON pixels.

Claim 66

The method as recited in claim 56 wherein the step of substantially eliminating horizontal lines from said image comprises the steps of: a) closing said image containing text and line graphics, said step of closing using a structuring element having at least two adjacent vertical ON pixels; and b) opening said image, said step of opening using a structuring element having at least two adjacent vertical ON pixels to produce said first intermediate image.

Claim 67

The method as recited in claim 57 preceded by the steps of: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 68

The method as recited in claim 58 preceded by the steps of: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 69

A digital computer system programmed to identify a text region in an image containing at least text and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 70

A digital computer system programmed to identify a text region in an image containing at least text and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, said at least a portion comprising said text region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of vertical lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a text image, said text image comprising a region of said image containing text and line graphics within said mask.

Claim 71

The computer system as recited in claim 61 programmed to eliminate vertical lines from said first intermediate image by: a) closing said image containing text and line graphics with a structuring element having at least two adjacent horizontal ON pixels; and b) opening said image with a structuring element having at least two adjacent horizontal ON pixels to produce said first intermediate image.

Claim 72

The computer system as recited in claim 62 wherein the structuring element used to open has at least three adjacent horizontal ON pixels.

Claim 73

The computer system as recited in claim 62 wherein the structuring element used to close has at least three adjacent horizontal ON pixels and the structuring element used to open has at least one more adjacent horizontal ON pixel than said structuring element used to close.

Claim 74

The computer system as recited in claim 62 further programmed to perform at least one reduction of said image containing text and line graphics.

Claim 75

The computer system as recited in claim 62 further programmed to perform at least two reductions of said image containing text and line graphics.

Claim 76

The computer system as recited in claim 61 programmed to turn on additional pixels by turning pixels on up to a bounding box.

Claim 77

The computer system as recited in claim 67 programmed to turn on additional pixels by iteratively eroding and dilating said first intermediate image with a structuring element having at least two adjacent diagonal ON pixels to produce a filled image until said filled image does not substantially change.

Claim 78

The computer system as recited in claim 67 further programmed to perform an additional reduction before turning on additional pixels.

Claim 79

The computer system as recited in claim 67 further programmed to perform a closing step, said closing step using a structuring element (SE) having at least three adjacent vertical ON pixels.

Claim 80

The computer system as recited in claim 67 further programmed to perform an opening step, said open using at least a 2×2 structuring element (SE) of solid ON pixels.

Claim 81

The computer system as recited in claim 67 further programmed to expand said mask by a factor of at least two.

Claim 82

The computer system as recited in claim 67 further programmed to dilate said mask.

Claim 83

The computer system as recited in claim 67 further programmed to: a) expand said mask then by a factor of at least two; b) dilate said mask then; c) again expand said mask by a factor of at least two.

Claim 84

A digital computer system programmed to identify a line graphics region in an image containing at least text and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 85

A digital computer system programmed to identify a line graphics region in an image containing at least text and line graphics, said digital computer system programmed to: a) convert OFF pixels adjacent text pixels to ON pixels using a structuring element having a higher probability of having a hit in said text than said line graphics, at least a portion of said ON pixels connecting adjacent text pixels so as to produce coalesced regions of ON pixels; and b) identify at least a portion of said image having said coalesced regions of ON pixels, at least a portion of a remainder of said image comprising said line graphics region, said identifying comprising: i) forming a mask, said mask comprising ON pixels in a region of text pixels, by substantially eliminating at least a portion of horizontal lines from said image containing text and line graphics to produce a first intermediate image, and turning on additional pixels adjacent to remaining ON pixels in said first intermediate image; and ii) forming a graphics image, said graphics image comprising a region of said image containing text and line graphics outside of said mask.

Claim 86

The computer system as recited in claim 75 programmed to substantially eliminate horizontal lines from said image by opening said image containing text and line graphics with a structuring element having at least two adjacent vertical ON pixels.

Claim 87

The computer system as recited in claim 75 programmed to substantially eliminate horizontal lines from said image by: a) closing said image containing text and line graphics, said step of closing using a structuring element having at least two adjacent vertical ON pixels; and b) opening said image, said step of opening using a structuring element having at least two adjacent vertical ON pixels to produce said first intermediate image.

Claim 88

The computer system as recited in claim 76 further programmed to process the image by: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.

Claim 89

The computer system as recited in claim 77 further programmed to process the image by: a) reducing said image containing text and line graphics to produce a reduced image; and b) closing said reduced image with a structuring element having at least five horizontal ON pixels.