Image processor with improved discrimination between character image and tonal image

Claim 2

An image processor comprising: discrimination means for discriminating image signals and for generating per pixel one of a character signal representing a character image and a tonal signal representing a tonal image, said tonal signal being successively generated in accordance with a length of said tonal image, whereby a series of tonal signals are generated; measuring means for measuring said length of said tonal image by counting a number of signals in said series of tonal signals. comparison means for comparing said number of signals with a first reference value; and converting means for converting said series of tonal signals to a series of character signals when said number of signals is smaller than said first reference value.

Claim 3

The image processor as set forth in claim 1, wherein said discrimination means comprises: diminishing means for diminishing a high-frequency component of said image signals; first classification means for comparing density data of an output signal from said diminishing means with a second reference value and for generating a binary signal pixel by pixel successively, whereby a series of binary signals including a binary signal for a target pixel is generated; and second classification means for generating either said character signal or said tonal signal for said target pixel with reference to said binary signal for a target pixel and binary signals for pixels in a region adjoining said target pixel.

Claim 4

The image processor as set forth in claim 2, wherein said second classification means comprises means for processing said binary signal and for generating said tonal signal only when all of said binary signals corresponding to said target pixel and said pixels in said region are at a high level.

Claim 5

The image processor as set forth in claim 3, wherein said second classification means has a crossed window sensing pattern extending in main and subordinative scanning directions from said target pixel by an identical number of pixels.

Claim 6

The image processor as set forth in claim 4, wherein the size of said crossed window sensing pattern is seven pixels times seven picture elements.

Claim 7

The image processor as set forth in claim 2, wherein said diminishing means comprises a low-pass filter which is substantially a crossed convolutional filter having an identical number of pixels extending from a target pixel in main and subordinative scanning directions.

Claim 8

The image processor as set forth in claim 6, wherein said convolutional filter causes the density value obtained from a division of a sum of the density values of image signals corresponding to all pixels inside of said convolutional filter by the number of pixels inside of said convolutional filter to turn into the density value of said target pixel.

Claim 9

The image processor as set forth in claim 6, wherein the size of said convolutional filter is three pixels times three pixels.

Claim 10

The image processor as set forth in claim 2, further comprising detection means for detecting a target pixel as an isolated image in a case in which all pixels surrounding said target pixel except for said target pixel are below a predetermined density value after the receipt of output signals from said diminishing means.

Claim 11

The image processor as set forth in claim 1, wherein said measuring means comprises a counter which is set by said tonal signal, said counter counting up synchronous with a clock signal and being reset by said character signal.

Claim 12

The image processor as set forth in claim 2, further comprising third classification means for classifying each of said image signals into a photographic image and others before the high-frequency component of said image signals is diminished by said diminishing means.

Claim 13

The image processor as set forth in claim 11, wherein said third classification means classifies each of said image signals per pixel and discerns that each of said image signals corresponding to a target pixel is substantially a photographic image responsive to a density depthwise difference between said target pixel and adjoining pixels being below a predetermined threshold value.

Claim 14

An image processor comprising: processing means having first multi-value coding means and second multi-value coding means for coding density data of image signals to multi-values; diminishing means for diminishing a high-frequency component of said image signals; first classification means for comparing density data of an output signal from said diminishing means with a second reference value and for generating a binary signal pixel by pixel successively, whereby a series of binary signals including a binary signal for a target pixel is generated; second classification means for generating one of a character signal representing a character image and a tonal signal representing a tonal image for said target pixel with reference to said binary signal for a target pixel and binary signals for pixels in a region adjoining said target pixel, said tonal signal being successively generated in accordance with a length of said tonal image, whereby a series of tonal signals are generated; measuring means for measuring said length of said tonal image by counting a number of signals in said series of tonal signals; comparison means for comparing said number of signals with a first reference value; converting means for converting said series of tonal signals to character signals when said counted number of signals is smaller than said first reference value; and selection means for, based on an output signal from said converting means, selecting either said first multi-value coding means or second multi-value coding means.

Claim 15

An image processor comprising: processing means having first multi-value coding means and second multi-value coding means for coding density data of image signals to multi-values; diminishing means for diminishing a high-frequency component of said image signals; first classification means for comparing density data of an output signal from said diminishing means with a second reference value and for generating a binary signal pixel by pixel successively, whereby a series of binary signals including a binary signal for a target pixel is generated; second classification means for generating one of a character signal representing a character image and a tonal signal representing a tonal image for said target pixel with reference to said binary signal for a target pixel and binary signals for pixels in a region adjoining said target pixel, said tonal signal being successively generated in accordance with a length of said tonal image, whereby a series of tonal signals are generated; measuring means for measuring said length of said tonal image by counting a number of signals in said series of tonal signals; comparison means for comparing said number of signals with a first reference value; converting means for converting said series of tonal signals to character signals when said counted number of signals is smaller than said first reference value; and selection means for, based on an output signal from said converting means, selecting either said first multi-value coding means or second multi-value coding means.

Claim 16

The image processor as set forth in claim 13, wherein said output signal from said diminishing means is also input to said processing means so that said output signal from said diminishing means is also input to said processing means so that said output signal from said diminishing means can be used for an image signal of the tonal image.

Claim 17

The image processor as set forth in claim 13, further comprising third classification means for classifying each of image signals into a photographic image and others before high-frequency components of said image signals are diminished by said diminishing means, and wherein, based on a signal output front said third classification means, a classified image signal of a photographic image is extracted and delivered to said processing means, and wherein the content of said multi-value coding process is selected based on signals output from said converting means and said third classification means.

Claim 18

The image processor as set forth in claim 15, wherein said output signal from said diminishing means is also delivered to said processing means so that said output signal from said diminishing means can be used for an image signal of a dot-line image.

Claim 19

The image processor as set forth in claim 13, wherein one of said first and second multi-value coding means includes means for binarizing said density data.