Photographic picture playback apparatus giving digital half-tone to screen picture facilitating adjustable roughness

[Detailed Description of the Invention]

[0001]

[Industrial Application]

The present invention relates to reproduction of a photograph picture, and although it can be made either black or white at every point in detail, it is related with generating electronically the screen (halftone dot) display picture which cannot be made into a middle gray level. Such a device contains the heat transfer fax machine, the laser electrostatic printer, and the ink-jet printer. The present invention relates to adjusting further the coarseness of the screen picture which made it produce.

[0002]

[Description of the Prior Art]

The usual electronic device for producing a photograph picture consists of a scanning module, a screen module, and a marking module. A scanning module detects the gray shade and shadow of each point of the original photograph picture, and it is used in order to carry out the electronic signal output of the information. A screen module processes this data so that it may become a desirable form for marking. Since many marking apparatus can only reproduce not gray middle shade and shadow but black, or white at the given arbitrary points, the screen module needs to generate the screen picture include only the point of black and white. Next, the electronic signal showing a screen picture is turned to a marking module. This marking module carries out marking of a medium like paper or a photograph film in respect of the black corresponding to the picture which occurs by a screen module, and white.

[0003]

One art used for a screen module is an electronic simulation which is the conventional screen art. This prior art is indicated to U.S. Pat. No. 498,127 (name "Screen For Making Photomechanical Printing Plates" of an invention "screen for carrying out making of the optical machine printing board"). The electronic simulation which is the present art is indicated to U.S. Pat. No. 4,012,584. This art simulates gray shade and shadow by changing the size of a dot. However, the number and position of these dots are still constant. When using it with a marking module with low surface imagery, this art produces two problems. The 1st problem is that a screen pattern is coarse. The 2nd problem is that there are few shade and shadow which may be distinguished. For these reasons, degradation like a recycled article arises. The conventional technology which changes the size of a dot does not use the accommodative or recursive method.

[0004]

The conventional screen art was used with the repetitive cell of the pixel (pixel) in a fixed position. The cell of these fixation is made to rotate occasionally in order to reinforce a pattern for those who see. Since a fixed cell has only a limited number of pixels, a restricted number of gray shade and shadow can only be obtained. For example, when there are 13 pixels in one cell, only the number of gray shade and shadow of 13 is possible. This is because each of 13 pixels must be either ON or OFF.

[0005]

Other art which may be used by a screen module is accommodative dithers (dither). According to this art, gray shade and shadow (intermediate color shade and shadow) are simulated using a very small dot. Brighter shade and shadow are denoted by dots fewer than darker shade and shadow. As an example in early stages of this art, it is given by U.S. Pat. No. 1,790,722 (name "Duplex Photomodulator" of an invention "duplex optical modulator"). As other general examples, Floyd (Floyd) R.W. and L. Staines Berg (Steinberg) -- work "An Adaptive Algorithm ForSpatial Grayscale" "accommodative algorithm for spatial gray gradation" Proc.SDI magazine vol.17 / the 2 75 to 77th pages, It is shown in the 279 to 283rd pages of "Digital Hlftoning" written by melon Tunei (Ulichney) R. "formation of digital half-tone." It is going to make the error signal showing the difference between a screen output and an input accommodative art approach zero. Usually, accommodative art generates the screen picture constituted from a dot of much very small uniform sizes.

[0006]

Although this accommodative art gives occasionally few patterns and the gray shade and shadow in which many distinction is more possible of detailed reproduction much more better than the conventional screen art and disturbance, The screen pattern generated is too fine not much, and it has another fault that it must have been reproduced well, in almost all marking apparatus. Especially when this problem reproduces the gray shade and shadow of 50% of gray area, it is serious.

[0007]

As mentioned above, with the present art, it is not suggested at all about changing the size of the dot which occurs with accommodative art.

[0008]

[Problem(s) to be Solved by the Invention]

The 1st issue that the present invention tends to solve in view of restriction of the pertinent art mentioned above is providing a photograph image restoration device with which a photograph picture is reproduced in the state of good detailed reappearance.

[0009]

The 2nd subject of the present invention is providing the photograph playback equipment whose distinction of much gray shade and shadow is enabled.

[0010]

The 3rd subject of the present invention is providing photograph playback equipment which a screen pattern produces preferably for eyes.

[0011]

The 4th subject of the present invention is providing the photograph playback equipment which a screen pattern can reproduce correctly with the marking apparatus of an ordinary form.

[0012]

There is the 5th subject of the present invention in changing the size of the dot which occurs with the accommodative art in photograph reproduction.

[0013]

[Means for Achieving the Goal]

The object which the present invention mentioned above is attained by recursive, i.e., a screen method which enables use of accommodative art, and device in order to change size of a dot used when creating a picture. The dot should care about being obtained by what (for example, white) marking of the medium which memorizes a created picture is carried out (for example, black), and is not done for marking. Here, marking means printing etc. Arbitrary print media which can express a picture with a memory value, an electronic medium, or other storages are within the limits of the present invention. Although a dot of black and white is meant, a dot of other colors may also be used within the limits of the present invention. Irregular arrangement of an arising dot improves greatly the number of gray shade and shadow which can be obtained by a marking apparatus. This recursive art operates in two dimensions, and uses a "hysteresis" or a coarseness usual state number which can be adjusted. This coarseness usual state number is used with an error signal, in order to determine coarseness of a picture.

[0014]

An image restoration device by the present invention has a screening (halftone-dot-izing) device which uses the RAM for random access memory (RAM), a scanning (scan) device, an input and an output, and temporary memory of an intermediate result, and a marking apparatus. It circulates through a screening device over three phases of operation, i.e., an operating state. Image data from a scanning device is memorized by input storage of RAM in the 1st operating state. Between the 2nd operating state, a screen pattern is calculated and it memorizes in an output region of RAM. A screen pattern memorized between the 3rd operating state in an output region of RAM is outputted to a marking apparatus. Each of these operating states is repeated over same number as a pixel (pixel) which exists in one scanning line of data elements. In order to avoid a disturbance diagonal line pattern between screen treatment operating states in every other one, subsequently to the left from the right, a scan is performed from the left on the right.

[0015]

For a person skilled in the art, clearly, processing for a screening device is performed using a software program which is general or runs by computer of a special purpose, or may be performed using an electronic circuit of a special purpose. Such both processings are within the limits of the present invention.

[0016]

[Example]

Although the present invention is applied to two-dimensional (x, y) screen treatment (halftone dot-ized processing), the case of one-dimensional (x) is assumed in order to simplify illustration. The input from a scanning device is expressed as i(x). i(x) is 256. It is assumed that it is a fraction between 0 and 1 which has a possible value of Pieces. In this case, each value expresses one of the gray shade and shadow of 256 (for example, 0 corresponds to white and 255 corresponds to black). For convenience, gray shade and shadow may be expressed by the integral value of 256 from zero (white) to total scale value fs (black) of 255. Of course, the display of black and white is optional and it can make it reverse. It is recognized as a marking apparatus producing only the dot of black and white, It is assumed that output O(x) of a screening device will be set to 0 if a marking apparatus must produce a white dot at point x with arbitrary points x, and it will be set to 1 if a marking apparatus must produce a black dot at point x. If you are trying for i(x) to serve as an integer between 0 and 255, O(x)=0 corresponds to i(x)=0 and O(x)=1 corresponds to i(x)=255. A scanning device needs to determine the size and the number of a dot of black required since the gray shade and shadow of the request which it reads in i(x) since i(x) can have a gray shade-and-shadow value of any value between 0 and 255 are produced, and whites. Dots by which marking is carried out on a medium may be called "Describing (indicia)" for convenience.

[0017]

The number and size of the black for the given gray shade and shadow and a white dot may be determined according to the present invention from the recursive relation in consideration of a result with the secondary error which are a pre- output, a front error, and a function of the present input. It is assumed that all the scales are expressed that fs mentioned above, and e(x) expresses the error between input i(x) and output O(x). In order to obtain the gray shade and shadow in an output equal to the shade and shadow in an input, it is wished for the average of the difference between input i(x) and output O(x), i.e., an error, to approach zero. Since i(x) can take the arbitrary integers between 0 and 255 and O(x) can take only discrete values 0 and 1, the mathematical relationship between i(x) and O(x) needs to include scaling for O(x). Therefore, at the time of an average, the difference between the products of fs showing i(x), O(x), and all the scales must approach zero, in order to obtain the result of a request. e(x) assumes that an error is defined and e'(x) defines a secondary error like the following formulas.

[0018]

The following formulas will be given if it is e(x) = e'(x) - fs * O(x)e'(x) = e(x-1)+i(x), next x= 3.

[0019]

Here, e (2) becomes like the following formulas.

[0020]

Therefore, the following formulas are obtained.

[0021]

e (3) = e (1) +i (2) - fs*O (2) +i(3)- fs*O (3) Here, e (1) becomes the following formulas.

[0022]

Therefore, e (3) is denoted by the following formulas.

[0023]

e (3) = e (0) +i (1) - fs*O (1) +i(2)- fs*O (2) +i (3) - fs*O (3) Here, e (0) is given by the following formulas.

[0024]

If referred to as e(-1) = 0, e (0) will become like the following formulas.

[0025]

e (0) = i (0) - fs * O (0), therefore e (3) become the following formulas.

[0026]

e(3) = i(0) - fs*O(0) + i(1) - fs*O(1) + i(2) - fs*O(2) + i(3) - fs*O (3), as a result the following expression 1 are obtained.

[0027]

[Equation 1]

If this is generalized, the following expression 2 will be given.

[0028]

[Equation 2]

Here, output O(x) from a screening device is defined like the following formulas.

[0029]

case [ ] of e'(x)+h * O(x-1) >= 0 -- O(x) = 1e' -- (x)+h * O (x-1) < -- 0 a case -- O(x) = 0 -- here, h is a hysteresis usual state number which defines, the range, i.e., the excursion, which can permit the surrounding error of zero.

[0030]

Therefore, as defined by the top, O(x) is a function of secondary error e'(x) and pre- output O (x-1). Secondary error e'(x) is a function of front error e (x-1) and input i(x). As shown in the top, front errors are the sum between input i(x) and output O(x), and a function of a difference. If e(x) is maintained by the range of general zero as shown in the expression of the upper sum, output O(x) is approximated to input i(x). A hysteresis usual state number expresses the difference between an output and an input in order to define the excursion which can permit the error near the zero. The arbitrary average errors which carried out division of e(x) by the number of the points to many points extremely are approximated to zero. Thus, the gray shade and shadow of an output become almost equal to the gray shade and shadow of an input.

[0031]

The influence of hysteresis usual state number h is denoted by the plot of e(x) opposite x shown in Drawings 1 and 2. If the gray shade-and-shadow input from a scanning device is regularity, i.e., a usual state number, because of simplification, all the values of i(x) become equal. in Drawing 1 -- the 1st usual state number gray shade and shadow -- 0 -- and an inclination wave-like error curve is produced among -h. This positive slope of an inclination waveform is i(x), and the dot of single white is made to be generated to all the x (when e(x) is between 0 and -h). For this reason, if this positive slope of the inclination waveform is including three points in the x direction, the dot of the single white of size 3 is generated. If e(x) reaches zero, the sign of inequality of the expression for O(x) will be reversed, and a black dot with single size corresponding to the number of the points of the x direction scanned by the negative slope of a curve will be generated. This negative slope of an inclination waveform between generating of a black dot is i(x)-1. Please care about that the slope of an inclination waveform is that of gray shade and shadow. Therefore, a positive and negative slope becomes equal to white and the middle fixed gray shade and shadow between black. The positive slopes of an inclination waveform decrease in number, and make a whiter pixel (pixel) output to brighter fixed gray shade and shadow. For a person skilled in the art, clearly, an above-mentioned direction and slope are arbitrary, and may be reversed within the limits of the present invention.

[0032]

Drawing 2 shows the influence of [ in case hysteresis usual state number h changes ]. Drawing 2 indicates that it assumed that increase of the value of h produces increase of the size of white and a black dot in Drawing 1 to the gray shade and shadow of the same usual state number. When h increases, it is because much more movement of an x-axis is needed in order that it may go up and down an inclination waveform. Therefore, as there is little transition to a white dot from a black dot, the arising screen picture becomes coarser.

[0033]

Although the continuous curve shows Drawings 1 and 2, Drawing 3 shows more correctly the case of a separate step at which point x is read separately and processed. This quantization is a function of a scanning device and does not change the principle of the present invention mentioned above.

[0034]

So far, the present invention was explained in relation to the dimension single for the object of illustrating easily. In order to expand to the usual two-dimensional half-tone (halftone), it is necessary to tackle distribution of the scanning technique and an error. One approach of a scan is considering it as the raster of the Serpentine (serpentine) form shown in Drawing 4. In that case, the scan of x (namely, pixel), i.e., a horizontal point, is made by the beginning from the left on the right, and then a left scan is performed from the right by y of a frame, i.e., the following line which fell perpendicularly. Therefore, the scan from the right and the left is every other one between perpendicular lines.

[0035]

One approach distributed in an error is assigning a half with error in each scanning direction. Therefore, it becomes a point (x-d, y) immediately preceded with a x direction to the 2-dimensional [ each ] point (x, y) scanned (being here). d is [ Is / Whether the direction of the present scan is the right from the left, or / it the left from the right? ] +1, or is -1 -- it becomes a point (y-x1) just above, and makes it each have a half with error shared

[0036]

Therefore, in a two-dimensional case, a recursive relation like a following formula is applied. e'(x, y) = (e(x-d, y) + e(x, y-1))/2 + i(x, y)(e(x-d, y) + h*(o(x-d, y) +o (y-x1) Case of >= 0 o(x, y) = 1 -- case it is other o(x, y) = 0e(x, y) = e'(x, y) - fs * o(x, y) d =-1y = 1 - 2 * (y mod 2) [0037]

Here, x is the point or pixel (pixel) of a scan line, and y is a scan line, i (x, y) and O (x, y) are an input and an output array, respectively, fs is a total scale input value which is usually 255 in the case of an 8-bit array, and h is usually 0.5 of the value of fs. It is a coarseness value which can adjust a twice as many value as this, and if the value of h becomes large, a screen will become coarse so much.

[0038]

The various scanning methods and error distribution are possible, and it can obtain corresponding expression from these scans and the error distribution method of entering within the limits of the present invention so that clearly [ a person skilled in the art ]. In addition, some approaches for constituting the screen method according to the present invention are possible. These include the software program memorized and run by computer of the electronic circuit designed especially in order to perform the method of the present invention, a special purpose, and the general object.

[0039]

Drawing 5 is a block diagram of the suitable example of the image restoration device using the method according to the present invention, for example, a photograph image restoration device. This photograph playback equipment measures the gray shade and shadow of the point of a picture, and contains scanning device 20 which changes these measured value into digital format (for example, 8-bit digital word showing the integer to 0 to 255), and transmits these to a screening device. Screening device 22 processes these data and generates the screen pattern which merely has two output possible-ized signals (others correspond to white corresponding to black in one). As mentioned above, these signals can respond to carrying out marking of black and the white middle, and not carrying out marking, respectively. In addition, arbitrary marking colors can be used. any of these examples -- although -- it is within the limits of the present invention. These signals from a screening device are transmitted to marking apparatus 24. This device carries out marking of a medium like one piece of paper, or a photograph film. For convenience, in the composition explained below, a screening device reads all the points in one line of a picture, i.e., the sequence of a pixel, calculates an error required after that, before it moves to the next line of a picture, it generates a signal, and it outputs this to a marking apparatus. Therefore, in the flow chart of Drawing 8 and Drawing 8A, the expression which defines the point of an array shows only one subscript x. Since two dimensions are expressed, 2nd subscript y used with the expression shown above is explained by line versus line scan. For example, it means that e(x) of a flow chart memorizes the error value only over one line y of array e (x, y) at once.

[0040]

Drawing 6 shows what realized the screening device by the present invention in the electronic circuit of the special object. This circuit contains output register 54 error register 32, output register 60, and in the past input register 28, error register 46, and in the past. The screen treatment device operates the following sequences in the screen treatment operating state over each pixel (pixel) of a picture.

[0041]

(1) The contents of error RAM30 corresponding to the present pixel are read into error register 32 in the past.

[0042]

(2) The contents of output RAM48 corresponding to the present pixel are read into output register 54 in the past.

[0043]

(3) The contents of input RAM26 corresponding to the present pixel are read into input register 28.

[0044]

(4) Both the contents of error register 32 are added, by shifting only 1 bit position to the right with counting-down circuit 36, are Divide(ed) to one half and, subsequently to the contents of input register 28, are added error register 46 and in the past.

[0045]

(5) Both the contents of output register 54 are added with adding machine 56, and multiply by them in multiplier 58 output register 60 and in the past, the coarseness value, i.e., hysteresis usual state number h, which can be adjusted from standard value generator 62.

[0046]

(6) Both a step (4) and the result of (5) are added. If this result is larger than zero or equal to it, the following steps (7a) will be performed. Otherwise, the following steps (7b) are performed.

[0047]

(7a) Memorize one usual state to output register 60 corresponding to the marking apparatus which carries out marking of the black dot of the position corresponding to the present pixel. A usual state number equal to the value of the input corresponding to black subtracts from the result of a step (4), and is memorized by error register 46. Usually, since eight bits expressed an input value and it was used, this usual state number subtracted 1 from the 8th power of 2 (28 - 1 =). It is 255.

[0048]

(7b) Memorize zero usual state corresponding to the dot of the white of the position corresponding to the present pixel to output register 60. The result of a step (4) is memorized by error register 46.

[0049]

(8) The contents of output register 60 corresponding to the present pixel are memorized by output RAM48.

[0050]

(9) The contents of error register 46 corresponding to the present pixel are memorized by error RAM30.

[0051]

(10) You are made to repeat this processing to the following pixel which becomes either the left of the present pixel or the right by the scanning direction.

[0052]

As mentioned above, hysteresis usual state number h influences the coarseness of the pattern formed with a screening device by moving the size of the dot made by a marking apparatus. 0 A of hysteresis usual state numerical value corresponds to the screen which can be made the finest. One of total scale value fs A twice as many hysteresis usual state numerical value as this corresponds to a coarse screen. To the screen of middle coarseness, it is 0.5 of total scale value fs. Twice is the usual value. However, h should care about that it can be made any value which is not smaller than 0.

[0053]

Drawing 6 is a figure showing connection of the circuit which performs the step of operation mentioned above. As shown in Drawing 6, screening device 22 has connection with scanning device 20 with which a data value is written in input RAM26. Although input RAM26, error RAM30, and output RAM48 are shown as a separate element, they can consist of portions of a single storage device, so that Well-known [ a person skilled in the art ]. Connection of input RAM26 others is used in order to write in the data value read from RAM26 to input register 28. As shown to the single line between input RAM26 and input register 28 by B, the trigger of this operation is carried out by B output of sequencer 64.

[0054]

Error RAM30 has connection with error register 32 in the past, and this is used in order to write the data value read from error RAM30 in error register 32 in the past. As indicated in the single line between error registers 32 as error RAM30 by B in the past, the trigger of this operation is carried out by B output of sequencer 64. It is used in order that other additional connection may memorize the contents of error register 46 to error RAM30. As shown to the path cord between error RAM30 and error register 46 by D, the trigger of this operation is carried out by output D of sequencer 64.

[0055]

Both the contents of register 46 are added by adding machine 34 the contents of error register 32, and in the past in the past. The result of this operation is Divide(ed) to one half with counting-down circuit 36. One example which Divide to one half is shifting only 1 bit position to the right. It will become more convenient, supposing it will also be possible to use other means to Divide to one half, for example, calculation will be made by a floating point operation Well-known, if it is a person skilled in the art. Such approach is a thing of the present invention within the limits. The result of this operation is added to the contents of input register 28 by adding machine 38. This result is added to other three circuits, such as usual state number subtraction machine 40, switch 44, and adding machine 50.

[0056]

Output RAM48 has connection with output register 54 in the past which is used in order to write the data value read from output register 48 in the past in output register 54 in the past. As shown to the signal wire between output registers 54 by B output RAM48 and in the past, the trigger of this operation is carried out by B output of sequencer 64. Additional connection is used in order to memorize the contents of output register 60 to output RAM48. As shown by D on the path cord between output RAM48 and output register 60, the trigger of this operation is carried out by sequencer 64. The data value to which another connection was read from output RAM48 is sent to a marking apparatus. As shown to the signal wire from output RAM48 to a marking apparatus by E, the trigger of this operation is carried out by output E of sequencer 64.

[0057]

The contents of output register 54 and the contents of output register 60 are mutually added by adding machine 56 in the past. Subsequently, it multiplies by this result by coarseness value 62 and multiplier 58 which can be adjusted. The result of this multiplication receives addition processing with adding machine 50. Then, this result has it examined from 0 with examination machine 52 whether it is size or it is equal to 0. Connection is made so that output C of sequencer 64 may enable it to memorize this result that is either 1 or 0 to output register 60.

[0058]

Therefore, output register 60 memorizes 1 here, when a marking apparatus needs to carry out marking of the black dot in the position corresponding to the present input pixel. A white dot must be given to the position corresponding to the present input pixel, then zero usual state is memorized by output register 60.

[0059]

Usual state number subtraction machine 40 memorizes a usual state number equal to the value of the corresponding input of the black which must be subtracted from an error register. Usually, since 8 bits expresses an input value and it is used, this usual state number is 8 of 2. 1 was subtracted from Power (28 - 1 =). It is 255.

[0060]

It is used in order that the result from examination machine 52 may switch between the outputs of adding machine 38 and usual state number subtraction machine 40, and the result subtracts 255 usual states from the result of adding machine 38 simply. If the result from examination machine 52 is 1, switch 42 closes, the trigger of the result of usual state number subtraction machine 40 will be carried out by output C of sequencer 64, and it will be memorized by error register 46 by it. If the result of examination machine 52 is 0, switch 44 will be closed and the result of adding machine 38 in which a trigger is again carried out by sequencer 64 will be memorized by error register 46.

[0061]

If sequencer 64 generates output D, the new value of output register 60 will be memorized by RAM48, and the new value of error register 46 will be memorized by RAM30. The whole processing is made to repeat here to the following pixel which can become either the right of the present pixel, or the left by whether the present line is a scan line of the number of even number or odd number.

[0062]

Drawing 7 expresses the signal made to be generated by the sequencer module. As mentioned above, sequencer 64 of Drawing 6 has a function which controls the sequence of the operation made by the circuit of a screening device. A sequence module generates the address signal for input RAM26, error RAM30, and output RAM48. A sequencer module has three operating states.

[0063]

Between the 1st operating state, to each pixel of the line concerned, sequencer module 64 generates output A, while counting an address from 0. This has the effect that the data value of one line from scanning device 20 is memorized by input RAM26.

[0064]

A sequencer module counts an address between the 2nd operating state to 0 to n-1, and n-1 to 0 between calls (invocations) of this 2nd operating state in every other one. Here, n is the number of pixels of each scan line. Between the counts of the 2nd operating state, a sequencer generates the pulse which are the output B, and C and D. Signal B has an effect which loads the value from corresponding RAM to registers 28, 32, and 54. Subsequently, the these-loaded value is processed by various circuits mentioned above since a new output and error value were generated. These new outputs and an error value will be memorized by output register 60 and error register 46 if sequencer 64 generates output C. As for output D of sequencer 64, finally, these memory values are memorized by output RAM48 and error RAM30, respectively.

[0065]

Between the 3rd operating state, sequencer 64 counts the address of a sequencer to 0 to n-1, while generating the E signal output to each pixel of the scan line concerned.

[0066]

Figure 8-8A is a flow chart which shows how it may be programmed, in order that the computer of the general object may function a screening device. The flow chart of figure 8-8A is a faithful simulation of the hardware shown in Drawing 6 and Drawing 7 so that clearly [ a person skilled in the art ]. This flow chart may comprise software which can run by computer of the general object.

[0067]

The flow chart of figure 8-8A shows again that the whole line is scanned with a scanning device, before screen treatment is performed. Therefore, as mentioned above, it is unnecessary that 2nd subscript y is shown in the above-mentioned expression. It is because this is coped with in the composition which operates by the whole scan line before moving to the following line. In functional block 801, the topmost line is denoted by line 0 and the scan to the right from the left is expressed by setting up variable dir equally to one. It is determined whether whether present value Y of the number of lines having been examined, and all the lines of the picture having been processed by functional block 803, and much more processing can be suspended. When it is denial, level variable X is set as 0 in functional block 805. In functional block 807, in order to determine whether all point i(x)s in line Y were read, the present value of X is examined. If it is denial, an input will be received from a scanning device, and measured value in(x) in point x will be memorized in functional block 811, and you will be made for X to be increased in functional block 809, as shown by block 813. These steps are repeated until the whole line will be read.

[0068]

Completion of reading of the line concerned will start screen treatment. The direction of a scan is examined as shown by block 815. When a direction directs a right scan from the left by 1, X is set as 0 and variable endX is set up equally to the total width of a frame. Otherwise, in block 816, X is set up equally to (all the width-1), and endX is set as -1. If the value of X is instructed not to be equal to endX and has not reached in this operating state of processing, as are shown by block 817, and the end of this scan line is shown by block 818, The recursive relation (here, the 2nd subscript is not needed) mentioned above is given in blocks 819 and 821. Functional block 821 shows temporary variable T showing the mean value which can be memorized by the register. In functional block 823, T is [ zero ] equal, or rather than it, when it is size, output O is set up equally to 1 and an error is set up equally to (error - 255). Here, 255 expresses the output of all the scales, i.e., black, as shown in functional block 825 and 827. Otherwise, as shown by functional block 829, output O is set up equally to 0. Variable T and arbitrary corresponding temporary memory registers may be omitted by unifying Step 821 and 823 to a single step so that clearly for a person skilled in the art. Next, in block 831, e(x) and o(x) are set up equally to an error and 0, respectively. You are made to increase X in block 833, and the processing is repeated until a line finishes.

[0069]

When processing of the line finishes, processing which gives the output signal over the line to a marking apparatus may be performed. Variable dir is made into a negative value as shown by block 835. Again, X is set up equally to 0 in block 837, and in block 839, it is examined in order to determine whether all the values over the scan line were told to the marking apparatus. If it is denial, in block 841, O will be set up equally to o(x), O will be outputted in block 843, and X will be increased in block 845. When an examination [ in / for a total of one scan line having been outputted to the marking apparatus / block 839 ] points, you are made to increase this scan line, as shown by block 847. As mentioned above, functional block 803 determines processing of the following line according to the same step until it finishes processing of all lines. Therefore, a scan, screening, and the output to a marking apparatus are completed before the following scan line is read.

[0070]

Although the present invention was mentioned above in relation to the specific example, it was only illustrated how a person skilled in the art would manufacture the present invention, and this would use it. Many examples of change are possible within the limits of the present invention. For example, recursive relations can be written using the assembler language for C program language and a specific computer, or other languages like FORTRAN (FORTRAN).

[0071]

[Effect of the Invention]

As mentioned above, in generating of the half-tone to which the screen of the photograph picture is carried out to the marking apparatus of a low resolution, with the conventional screen art, the display was too coarse, and it was too fine with accommodative dither art. In the present invention, the screen pattern which enabled regulation of coarseness is given, and the strong point of the conventional accommodative dither art is not spoiled. The present invention is a hysteresis usual state number and recursive art so that the size of the dot of a screen picture may be changed, It applies to the accommodative screen art restricted conventionally [ provided with the fixed dot pattern ], and a hysteresis usual state number is changed by it, regulation of the coarseness of a picture is enabled and the subject of an invention had and mentioned above can be solved effectively.

[Brief Description of the Drawings]

[Drawing 1]

It is a figure showing the fixed gray shade and shadow inputted from a scanner, corresponding black, and the error curve to a white output.

[Drawing 2]

They are Drawing 1 which has a hysteresis usual state number of a big value, and a figure showing the same curve.

[Drawing 3]

It is a figure showing the effect of the quantization which measures an individual point.

[Drawing 4]

It is a figure showing the scan by the curve type (serpentine) raster processing method.

[Drawing 5]

It is a block diagram of the suitable example of a photograph image restoration system.

[Drawing 6]

It is a block diagram of a screening device.

[Drawing 7]

It is a figure showing the table of the signal made to be generated by the sequencer module.

[Drawing 8]

In order that the computer of the general object may function a screening device, it is a flow chart which shows how it is programmed.

[Drawing 8A]

In order that the computer of the general object may function a screening device, it is a flow chart which shows how it is programmed.

[Description of Notations]

20 Scanning Device 22 Screening Device 24 Marking Apparatus 26 Input RAM 28 Input Register 30 Error RAM 32 Past Error Register 34 Adding Machine 36 Counting-down Circuit 40 Usual State Number Subtraction Machine 42 Switch 44 Switch 46 Error Register 48 Output RAM 50 Adding Machine 52 Examination Machine 54 Past Output Register 56 Adding Machine 58 Multiplier 60 Output Register 62 Usual State Number Generator 64 Sequencer