Edge extraction technique

Claim 1

A method of producing an enhancement value for enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, isolating an m bit by m bit segment of the raster, where m is an odd number, and the N bit is the center bit, calculating an edge sharpness value of the horizontal edges directly above and below the N bit, calculating an edge sharpness value of the vertical edges directly to the left and right of the N bit, for each direction, finding a most positive and a most negative sharpness value from the two sharpness values and zero, taking the root mean square of the two most positive values, and of the two most negative values, to create positive and negative enhancement values, testing the enhancement values by dividing their difference by the larger of the two, and comparing the absolute value of the result of said division to a first threshold, using the positive enhancement value as the enhancement value if N is a "1" and the first threshold was not exceeded, using the negative of the negative enhancement value as the enhancement value if N is a "0" and the first threshold was not exceeded, using the positive enhancement value as the enhancement value if the positive value is larger than the negative value, and if the first threshold was exceeded, selecting the negative of the negative enhancement value as the enhancement value if the negative value is larger than the positive value, and if the first threshold was exceeded, testing the absolute value of the selected enhancement value against a second threshold, adding the selected enhancement value to the sum of the "1" bits in said m by m bit segment to create an edge enhanced gray scale pixel if the second threshold is exceeded, or adding nothing to said sum if the second threshold is not exceeded.

Claim 2

A method of producing an enhancement value for enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, isolating an m bit by m bit segment of the raster, where m is an odd number, and the N bit is the center bit, calculating an edge sharpness value of the horizontal edges directly above and below the N bit, calculating an edge sharpness value of the vertical edges directly to the left and right of the N bit, for each direction, finding a most positive and a most negative sharpness value from the two sharpness values and zero, taking the root mean square of the two most positive values, and of the two most negative values, to create positive and negative enhancement values, testing the enhancement values by dividing their difference by the larger of the two, and comparing the absolute value of the result of said division to a first threshold, using the positive enhancement value as the enhancement value if N is a "1" and the first threshold was not exceeded, using the negative of the negative enhancement value as the enhancement value if N is a "0" and the first threshold was not exceeded, using the positive enhancement value as the enhancement value if the positive value is larger than the negative value, and if the first threshold was exceeded, selecting the negative of the negative enhancement value as the enhancement value if the negative value is larger than the positive value, and if the first threshold was exceeded, testing the absolute value of the selected enhancement value against a second threshold, adding the selected enhancement value to the sum of the "1" bits in said m by m bit segment to create an edge enhanced gray scale pixel if the second threshold is exceeded, or adding nothing to said sum if the second threshold is not exceeded.

Claim 3

A method of enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, isolating a seven bit by seven bit segment of raster, the N bit being at the center, determining the number of bits in each column for columns 1-7 and rows 1-7, where Cn is defined as the number of the column and Rn is defined as the number of the row in said segment, solving the following equations, where N0 is the sum of pixels in the segment, Ne is the gray level value of pixel N after enhancement, T1 is a first threshold, T2 is a second threshold, and a is a scaling coefficient, d1x=(C2+C3+C4)-(C5+C6+C7), d2x=(C6+C5+C4)-(C3+C2+C1), d1y=(R2+R3+R4)-(R5+R6+R7), d2y=(R6+R5+R4)-(R3+R2+R1), dxp(ositive)=MAX(0,d1x,d2x), dxn(egative)=MIN(0,d1x,d2x), dyp(ositive)=MAX(0,d1y,d2y), dyn(egative)=MIN(0,d1y,d2y), dxyp=SQRT(dxp.sup.2 +dyp.sup.2)×a, dxyn=SQRT(dxn.sup.2 +dyn.sup.2)×a, if ABS(dxyp-dxyn)/MAX(dxyp, dxyn)≦T.sub.2, e=dxyp, with N=1, e=-dxyn, with N=0, otherwise, e=dxyp, with dxyp greater than dxyn, e=dxyn, with dxyp less than dxyn, if ABS(e) is greater than T 1 a, N e =N o +e, otherwise, N e =N o .

Claim 4

A method of enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, isolating a seven bit by seven bit segment of raster, the N bit being at the center, determining the number of bits in each column for columns 1-7 and rows 1-7, where Cn is defined as the number of the column and Rn is defined as the number of the row in said segment, solving the following equations, where N0 is the sum of pixels in the segment, Ne is the gray level value of pixel N after enhancement, T1 is a first threshold, T2 is a second threshold, and a is a scaling coefficient, d1x=(C2+C3+C4)-(C5+C6+C7), d2x=(C6+C5+C4)-(C3+C2+C1), d1y=(R2+R3+R4)-(R5+R6+R7), d2y=(R6+R5+R4)-(R3+R2+R1), dxp(ositive)=MAX(0,d1x,d2x), dxn(egative)=MIN(0,d1x,d2x), dyp(ositive)=MAX(0,d1y,d2y), dyn(egative)=MIN(0,d1y,d2y), dxyp=SQRT(dxp.sup.2 +dyp.sup.2)×a, dxyn=SQRT(dxn.sup.2 +dyn.sup.2)×a, if ABS(dxyp-dxyn)/MAX(dxyp, dxyn)≦T.sub.2, e=dxyp, with N=1, e=-dxyn, with N=0, otherwise, e=dxyp, with dxyp greater than dxyn, e=dxyn, with dxyp less than dxyn, if ABS(e) is greater than T 1 a, N e =N o +e, otherwise, N e =N o .

Claim 5

A method of producing an enhancement value for enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, each bit being a 1 or a 0, isolating an m bit by m bit segment of the raster, where m is an odd number, and the N bit is the center bit, calulating an edge sharpness value of the horizontal edges directly above and below the N bit, calculating an edge sharpness value of the vertical edges diretly to the left and right of the N bit, for each direction, finding a most positive and a most negative sharpness value from the two sharpness values and zero, taking the root mean square of the most positive values, and of the two most negative values, to create positive and negative enhancement values, testing the enhancement values by dividing their difference by the larger of the two, and comparing the absolute value of said division to a first threshold, using the positive enhancement value as the enhancement value if N is a "1" and the first threshold was not exceeded, using the negative of the negative enhancement value as the enhancement value if N is a "0" and the rirst threshold was not exceeded, using the positive enhancement value as the enhancement value if the positive value is larger than the negative value, and if the first threshold was exceeded, selecting the negative of the negative enhancement value as the enhancement value if the negative value is larger than the positive value, and if the first threshold was exceeded, testing the absolute value of the selected enhancement value against a second threshold, adding the selected enhancement value to the sum of the "1" bits in said 7 by 7 segment to create an edge enhanced gray scale pixel if the second thresold is exceeded, or adding nothing to said sum if the second threshold is not exceeded.

Claim 6

A method of producing an enhancement value for enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, each bit being a 1 or a 0, isolating an m bit by m bit segment of the raster, where m is an odd number, and the N bit is the center bit, calulating an edge sharpness value of the horizontal edges directly above and below the N bit, calculating an edge sharpness value of the vertical edges diretly to the left and right of the N bit, for each direction, finding a most positive and a most negative sharpness value from the two sharpness values and zero, taking the root mean square of the most positive values, and of the two most negative values, to create positive and negative enhancement values, testing the enhancement values by dividing their difference by the larger of the two, and comparing the absolute value of said division to a first threshold, using the positive enhancement value as the enhancement value if N is a "1" and the first threshold was not exceeded, using the negative of the negative enhancement value as the enhancement value if N is a "0" and the rirst threshold was not exceeded, using the positive enhancement value as the enhancement value if the positive value is larger than the negative value, and if the first threshold was exceeded, selecting the negative of the negative enhancement value as the enhancement value if the negative value is larger than the positive value, and if the first threshold was exceeded, testing the absolute value of the selected enhancement value against a second threshold, adding the selected enhancement value to the sum of the "1" bits in said 7 by 7 segment to create an edge enhanced gray scale pixel if the second thresold is exceeded, or adding nothing to said sum if the second threshold is not exceeded.

Claim 7

A method of producing an enhancement value for enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, each bit being a 1 or a 0, isolating an m bit by m bit segment of the raster, and the N bit is the center bit, calulating an edge sharpness value of the horizontal edges directly above and below the N bit, calculating an edge sharpness value of the vertical edges diretly to the left and right of the N bit, for each direction, finding a most positive and a most negative sharpness value from the two sharpness values and zero, taking the root mean square of the most positive values, and of the two most negative values, to create positive and negative enhancement values, testing the enhancement values by dividing their difference by the larger of the two, and comparing the absolute value of said division to a first threshold, using the positive enhancement value as the enhancement value if N is a "1" and the first threshold was not exceeded, using the negative of the negative enhancement value as the enhancement value if N is a "0" and the rirst threshold was not exceeded, using the positive enhancement value as the enhancement value if the positive value is larger than the negative value, and if the first threshold was exceeded, selecting the negative of the negative enhancenent value as the enhancement value if the negative value is larger than the positive value, and if the first threshold was exceeded, testing the absolute value of the selected enhancement value against a second threshold, adding the selected enhancement value to the sum of the "1" bits in said segment to create an edge enhanced gray scale pixel if the second threshold is exceeded, or adding nothing to said sum if the second threshold is not exceeded, and wherein the root mean square values are multiplied by a factor, less than one, to allow an adjustment of the amount of extracted edge to be used for enhancement.

Claim 8

A method of producing an enhancement value for enhancing the edges of a halftone copy of a halftone original comprising the steps of: scanning the original to produce a one bit per pixel raster, each bit being a 1 or a 0, isolating an m bit by m bit segment of the raster, and the N bit is the center bit, calulating an edge sharpness value of the horizontal edges directly above and below the N bit, calculating an edge sharpness value of the vertical edges diretly to the left and right of the N bit, for each direction, finding a most positive and a most negative sharpness value from the two sharpness values and zero, taking the root mean square of the most positive values, and of the two most negative values, to create positive and negative enhancement values, testing the enhancement values by dividing their difference by the larger of the two, and comparing the absolute value of said division to a first threshold, using the positive enhancement value as the enhancement value if N is a "1" and the first threshold was not exceeded, using the negative of the negative enhancement value as the enhancement value if N is a "0" and the rirst threshold was not exceeded, using the positive enhancement value as the enhancement value if the positive value is larger than the negative value, and if the first threshold was exceeded, selecting the negative of the negative enhancenent value as the enhancement value if the negative value is larger than the positive value, and if the first threshold was exceeded, testing the absolute value of the selected enhancement value against a second threshold, adding the selected enhancement value to the sum of the "1" bits in said segment to create an edge enhanced gray scale pixel if the second threshold is exceeded, or adding nothing to said sum if the second threshold is not exceeded, and wherein the root mean square values are multiplied by a factor, less than one, to allow an adjustment of the amount of extracted edge to be used for enhancement.