JP2002152515A - Image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor in which a good image quality can be ensured by removing noise conveniently using a small number of memories. SOLUTION: A noise removing circuit is arranged such that multilevel data of 5 lines (41-45) is fed to a noise removing circuit 46 by FIFO(first in first out) system, a discrimination is made between an isolated point and a nonisolated point based on the results of a first decision means 47 and a second decision means 48 in the noise removing circuit 46, and decision results data 49 is delivered to the post-stage. A data area used for decision at the decision means 47 and 48 is constituted of an M×N pixel matrix including the area 51 of an interested pixel, and two L×K pixel areas (K<=L) adjacent to the matrix area on the opposite sides thereof. Size of a removable noise or removable conditions can be varied by setting the size of M, N, L and K arbitrarily.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus, and more particularly, to a printer, a digital PPC, and a facsimile.
More particularly, the present invention relates to an image processing apparatus that uses a plurality of devices that perform SIMD control on a plurality of element processors and determines a feature amount effective area of each element.

[0002]

2. Description of the Related Art With respect to image processing, a conventional method for removing noise isolated points is disclosed in Japanese Patent No. 2853140, "Image Area Identification Apparatus", JP-A-6-30265, "Image Processing Apparatus", and JP-A-6-30266. "Image processing device",
The methods described in JP-A-10-229497 "Image processing method, image processing apparatus and recording medium" are common, and they are generally classified into two methods.

A first method is disclosed in Japanese Patent No. 2853140.
JP-A-6-30265, JP-A-6-3026
In the invention described in Japanese Patent Application Laid-Open No. 6-260, a pattern is detected by binarizing image data, and a background and an isolated point are easily distinguished by binarizing a density level. As a second method, JP-A-10-229497
In the invention described in Japanese Patent Application Laid-Open Publication No. H10-209, global pixel information is calculated for binarized data.

[0004]

However, in the first method, if the binarized data is used to make it easy to distinguish the background from the isolated noise, the resolution of the image forming apparatus is increased. The number of constituent pixel data in the isolated point data composed of several pixels increases,
As a result, there is a problem that the boundary between the noise isolated point and the background cannot be detected unless the data for determination is viewed in a wider pixel range, that is, it is difficult to remove the noise isolated point. Furthermore, to solve this,
There is a problem that a line memory for viewing in a wide pixel range increases.

[0005] In the second method, if it is desired to use multi-valued image data in the process after the isolated noise removal process, the multi-valued image data is only as much as seen over a wide range. Another problem is that a memory for storing binarized data is required.

SUMMARY OF THE INVENTION The present invention has been made in view of the above situation, and extracts noise isolated points without increasing the reference pixel range and without storing binarized data separately from image data. It is another object of the present invention to provide an image processing apparatus which removes or makes the image inconspicuous, can easily remove noise with a small memory, and can obtain good image quality.

According to the present invention, the image data area surrounding the pixel of interest is not expanded by the resolution, and furthermore, the image data is kept multi-valued, and the character and picture data and the isolated points are distinguished with high accuracy. Perform noise isolated point removal with accuracy,
It is an object of the present invention to provide an image processing device capable of obtaining a good image.

Further, the present invention provides an image processing apparatus capable of distinguishing low-density halftone dot isolated points and noise isolated points with high accuracy without using binary data and with image data remaining multi-valued. Its purpose is to provide.

According to the present invention, there is provided an image processing apparatus capable of passing appropriate feature amount data to a subsequent stage process even when noise isolated points are removed after the feature amount calculating means for the pixel of interest. Its purpose is to provide.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image processing apparatus capable of making an unnecessary edge portion less noticeable at a noise isolated point when a target pixel is an isolated point.

[0011]

According to a first aspect of the present invention, it is determined whether each pixel data constituting image data is an isolated point or a non-isolated point, and the pixel data determined to be an isolated point is determined. An image processing apparatus for removing data, comprising: first determining whether or not a target pixel is a candidate for an isolated point from multi-valued data of M × N pixels (M and N are integers) including target pixel data; And L × K adjacent to the M × N pixel area
A second determination that corrects the result of the first determination unit from multi-valued data in a pixel area (L and K are integers; K ≦ N) and determines whether the target pixel is an isolated point. Means, and data noise removing means for removing the pixel data of interest when the pixel data of interest is an isolated point according to the result of the determination by the second determining means.

According to a second aspect of the present invention, in the first aspect of the present invention, there is further provided a feature amount correcting means for correcting feature amount data corresponding to the target pixel data removed by the data noise removing means. It was done.

According to a third aspect of the present invention, in the first or second aspect of the present invention, the first determining means determines a number of adjacent t (t is an integer) in the oldest line among the M × N pixels. A level value determining circuit for selecting a larger value as a level value between a sum of pixel data and a sum of data of adjacent t pixels in the newest line among the M × N pixels; A reference density extraction circuit that extracts a sum of data of adjacent t pixels including the target pixel as a reference density value;
When any or all of the difference in the vertical direction, the difference in the diagonal right direction, and the difference in the diagonal left direction are larger than a predetermined threshold value, the pixel of interest is determined to be a candidate for an isolated point. is there.

According to a fourth aspect of the present invention, in the first aspect of the invention, the second judging means includes the M
In the area of L × K pixels adjacent to the area of × N pixels, a continuous amount of pixels whose pixel data is smaller than a predetermined threshold is calculated, and when the continuous amount is smaller than a predetermined value, the first determination is made. Even if the result of the means is an isolated point candidate, the pixel of interest is determined to be a non-isolated point.

According to a fifth aspect of the present invention, in any one of the second to fourth aspects of the present invention, the feature amount correcting means determines that the target pixel is an isolated point by the first determining means and the second determining means. When it is determined that there is, the feature amount data is replaced with an arbitrary value.

According to a sixth aspect of the present invention, in the first aspect of the present invention, it is determined that the target pixel is an isolated point based on a result of the determination by the first determining means and the second determining means. In this case, the data of the pixel of interest is replaced with an arbitrary smoothed value.

[0017] The inventions of claims 7 and 8 correspond to claim 1 of the present invention.
A program for causing the computer to function as each means according to any one of the first to sixth aspects, and a computer-readable recording medium on which the program is recorded.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An image processing apparatus according to an embodiment of the present invention refers to multivalued data as a parameter to make a determination, so that an image data area surrounding a pixel of interest is not expanded according to resolution. The noise isolated points are removed with high accuracy while the data remains multi-valued, so that a good image is obtained.

This embodiment is based on SIMD (single instruc
The operation will be described in detail by taking an image processing unit of a digital copying machine that processes with a processor as an example.
FIGS. 1 to 4 are block diagrams each showing a configuration example of an image processing unit of a digital copying machine according to an image processing apparatus according to an embodiment of the present invention. In the example of the image processing unit illustrated in FIG. 1, input data from a scanner 1 as an input device is input to a noise removal circuit 3 via a shading circuit 2. The noise elimination circuit 3 removes unnecessary noise isolated points on the image data of the document, and thereafter, the feature quantity extraction circuit 4 extracts the feature quantity of the image.
The data is sent to the subsequent filter processing circuit 5.

In the example of the image processing unit shown in FIG. 2, the image data input from the scanner 11 is subjected to the shading circuit 12 and then the characteristic amount is extracted by the characteristic amount extracting circuit 14 first. Thereafter, noise is removed in the noise removing circuit 13. Therefore, since the data up to the shading circuit 12 and the data after the noise removal change, the feature amount extracted from the data after the shading does not match the data passed from the noise removal circuit 13 to the subsequent filter processing circuit 15. There is a part. Thus, the feature amount is corrected based on the result of the noise removal circuit 13.

In the example of the image processing unit shown in FIG. 3, the image data input from the scanner 21 is first passed through the shading circuit 22 and then the characteristic amount is extracted by the characteristic amount extracting circuit 24. The noise removal by the filter 23 is performed simultaneously with the filter calculation by the filter processing circuit 25. In this case, instead of performing a filter calculation on the data from which noise has been removed, the filter operation result can be removed or not removed according to the noise determination result. Although the filter operation itself is calculated using the image data with noise, the line memory required for noise removal and the line memory required for the filter operation can be shared. In this example, the noise removal is performed after the filter operation based on the feature amount result, so that feature amount correction is not required.

The image processing unit shown in FIG. 4 converts the image data input from the scanner 31 into a shading circuit 3.
2, the feature amount is extracted by the feature amount extraction circuit 34,
This is a configuration example in the case where the feature amount extraction circuit 34 and the noise removal circuit 33 share a line memory. The feature amount is extracted from the image data including the noise data, and the noise is removed from the image data arranged in the same line memory. Therefore, the feature amount needs to be corrected. Thereafter, the filter processing circuit 35 performs the filter processing.

FIG. 5 is a block diagram showing a configuration example of the noise elimination circuit according to one embodiment of the present invention with respect to any one of FIGS. As a noise removal circuit, for example,
The data of 5 lines (41-45) is stored in FIFO (first
The signal is input to the noise removing circuit 46 in an in first out) manner, and the result of the first determining means 47 and the second determining means 48 in the noise removing circuit 46 is used to determine whether the point is an isolated point or a non-isolated point. What is necessary is just to comprise so that the data 49 can be sent to the latter stage.

FIG. 6 is a diagram showing an example of a data area used for determination by the determination means in FIG. The data area used for the determination by the determination means 47 and 48 in FIG. 5 is an M × N pixel matrix including a target pixel area 51 and two L × K pixel areas 52 adjacent to and sandwiching the matrix area. Consists of FIG. 6 shows the case where K = N, but it is sufficient if K ≦ N, and M, N, L, K
By arbitrarily setting the size or the ratio thereof as a parameter, it is possible to change the size of noise that can be removed and the conditions under which noise can be removed.

FIG. 7 is a flowchart for explaining a determination method according to an embodiment of the present invention. In the determination of the noise isolated point, when the noise isolated point determination result (koritsu) is true (YES in step S1), the pixel data of interest is removed (step S3), otherwise, the data remains as it is (step S3). S2).

An image processing apparatus according to another embodiment of the present invention refers to multi-valued data of M × N pixels as a parameter to make a determination, so that an image data area surrounding a pixel of interest is not expanded according to resolution, and characters and characters are not expanded. The pattern data and the isolated points are identified with high accuracy. Regarding the image processing apparatus according to the present embodiment, only the parts added or changed to the above embodiment will be described below.

FIG. 8 is a block diagram showing an example of the internal configuration of the first determination means in FIG. In the present embodiment, the determination is made using the data of the oldest line 63, the data of the latest line 61, and the data of the target pixel line 62 among the line memories. Each pixel of each line data is added to an adjacent t pixel of the same line by an adder 6.
Add at 4. In the level value determination circuit 65, the addition data to the line memory 63 and the latest line memory 61
Is determined as the level value. In the reference density extraction circuit 66, the target pixel line 6
The addition data for 2 is extracted as reference density data. The subtracter 67 calculates (reference density data)-(level value)
Is calculated, and the result is calculated by a comparator 68 as a threshold (th).
Is determined by the determination unit 69 as an isolated point if it is equal to or larger than the threshold, and a non-isolated point if it is smaller than the threshold.

FIG. 9 is a block diagram showing another example of the internal configuration of the first determining means in FIG. FIG. 9 shows FIG.
The same operation as that described above is performed, but the data of lines 71, 72, and 73 are each added to the adder 74 by dividing by 1 / f, so that the data to be handled can be reduced and the calculation can be simplified. I have to. Each line data 71, 7
Of the 2,73, the oldest line 73 and the latest line 7
1 and the data of the target pixel line 72.
Each pixel of each line data is divided by 1 / f, and adjacent t pixels on the same line are added by the adder 74. The level value determination circuit 75 determines the larger of the added data for the line 73 and the added data for the latest line 71 as the level value. The reference density extraction circuit 76 extracts the added data for the target pixel line 72 as reference density data. The subtractor 77 calculates (reference density data)-(level value), and the result is compared with a threshold value (th) by a comparator 78. Non-isolated point if smaller than the threshold,
Is determined.

FIG. 10 shows a method of adding each line by SIMD.
FIG. 3 is a diagram schematically illustrating a configuration of a processor. When the pixel of interest is p, for example, both adjacent pixels p +
The three pixels of 1, p-1 are added. At this time, the target pixel p
, The same operation is performed for all pixels, and the addition result of three pixels is input to the register (shaded portion) of each pixel. The register of the target pixel p stores the three-pixel addition result P. Similarly, in the line 61, the register a of the pixel a stores the sum a of the pixel data a and the adjacent pixel data, and the same operation is performed in all the pixels. A + 2 is stored in the register of the pixel a-2, and the sum A-2 of the adjacent pixels is also stored in the register of the pixel a-2. The same operation is performed in the line 63.

FIG. 11 is a two-dimensional image of FIG. Now, assuming that the pixel of interest is p, the sum P of the adjacent pixels p−1 and p + 1, and A and A +
2, A-2 and B, B + 2, B-2 of the line 63 have a positional relationship as shown in the figure.

FIG. 12 shows A + 2, P, B shown in FIG.
FIG. 9 is a diagram for explaining a determination method using data of −2. The determination using the data of A + 2, P, and B-2 prevents the diagonal line rising to the right from being determined as an isolated point. The oblique line indicates the maximum value of A + 2 and B-2, and P
Is small, and if the difference is equal to or smaller than the threshold, it is determined to be a non-isolated point.

FIG. 13 is a view for explaining a determination method using the data of A, P, and B shown in FIG. A,
By using the data of P and B, a vertical thin line is not determined as an isolated point.

FIG. 14 shows A-2, P, B shown in FIG.
It is a figure for explaining the judgment method using +2 data. By using the data of A-2, P, and B + 2, the diagonally downward slant line is determined as a non-isolated point.

FIGS. 15A and 15B are diagrams for explaining a method of making a decision after binarization according to the prior art. FIG. 15A shows a pixel, and FIG. 15B shows the decision value. FIG.
In the conventional method illustrated in FIG.
In order to be converted to value data, in order to be determined as an isolated point, as shown in FIG. 15A, the data of the lines 61 and 63 are close to the background and are values that become white by binarization. The data 62 near the target pixel needs to be a value that becomes black when binarized. For this reason, a large size noise isolated point cannot be extracted. It is necessary to lower the threshold value for binarization as the noise becomes larger, so that erroneous determination that data becomes noise more than necessary is likely to occur.

FIG. 16 shows a judgment method according to the present invention in FIG.
16A is a diagram illustrating a pixel, and FIG. 16B is a diagram illustrating a determination value thereof. In the method according to the present invention illustrated in FIG. 16, the noise isolated point is large, and even if the data of the edge portion of the noise pixel is applied to the data of the lines 61 and 63, if the data value is dull, the reference density data and Can be determined as an isolated point by the difference of

An image processing apparatus according to another embodiment of the present invention refers to multi-valued data in an L × K pixel area as a parameter to make a determination, so that image data is multi-valued without using binarized data. The low density halftone dot isolated point and the noise isolated point are identified with high accuracy. Regarding the image processing apparatus according to the present embodiment, only the parts added or changed from the above embodiments will be described below.

FIG. 17 is a diagram showing an example of the area of the second determination means. In the present embodiment, the determination is performed in the areas on both sides adjacent to the area 81 of the first determination means as the area 82 used for the second determination means. Thus, a horizontal line that cannot be determined as a non-isolated point by the first determination unit can be set as a non-isolated point. Second determination area 8
If all of the pixel data of the second and third pixel data are lower than a predetermined threshold value (can be set arbitrarily), the first determination result is output as it is. If there is pixel data larger than an arbitrary threshold value in the second determination area 82, it is determined as a non-isolated point even if the first determination result is an isolated point candidate. The values of L and K are necessary to determine an isolated point under a certain condition as an isolated point.
It is determined by the continuous amount of pixel data smaller than the threshold.

FIG. 18 is a diagram for explaining the operation of the second judging means for judging in the second judging area of FIG. 17. In FIG. 18, D1 and D2 represent lines, respectively. For example, the image data d1, d2, d3 in the D1 line
Each of them is compared with the threshold value, and when the data is larger than the threshold value, the first bits of the flags F1 to F3 of the respective pixel columns 91, 92, and 93 are set. FIG. 18 shows a case where, on the D1 line, d1 and d3 are larger than the threshold, and only d2 is smaller than the threshold.

Similarly, each pixel on the line D2 is compared with the threshold value, and the result is reflected in the second bit of the flags F1 to F3 of the pixel rows 91, 92, and 93. As a result, it is understood that each pixel column has flag data of “01” in the pixel column 91, “10” in the pixel column 92, and “11” in the pixel column 93. When all of d1 to d6 shown in FIG.
Should be stored. The determination in the second determination area can be performed using the storage result of the flag.

FIG. 19 is a block diagram showing an example of a circuit for realizing the operation of FIG. Image data 101
On the other hand, when the sum of the flag results (for example, the pixel columns 91 to 93) described in FIG.
When the determination result of the determination means 103 is an isolated point, it is set to 0, and when it is a non-isolated point, it is set to 1 and the result of inverting the data by the inversion circuit 104 and the determination of the first determination means 102 Based on the result and the AND condition in the AND circuit 105, a final isolated point / non-isolated point determination is performed.

The image processing apparatus according to another embodiment of the present invention reflects the determined result on the characteristic amount data of the pixel of interest, so that a noise isolated point is located downstream of the characteristic amount calculating means for the pixel of interest. Even when the removal is performed, appropriate feature amount data is passed to the subsequent processing. Regarding the image processing apparatus according to the present embodiment, only the parts added or changed from the above embodiments will be described below.

FIGS. 20 and 21 are flow charts for explaining a judgment method according to another embodiment of the present invention. As in the embodiment described with reference to FIG. 2 and FIG.
As shown in FIG. 0 and FIG. 21, feature amount data, for example, binary edge data in FIG. 20, and multi-value edge amount data in FIG. 21 are corrected. In FIG. 20, since the data is binary edge data, and a noise isolated point is determined to be an edge, the data is inverted. That is, in the determination of the noise isolated point, if the noise isolated point determination result (koritsu) is true (YES in step S11), the edge data is inverted (step S13);
The data remains as it is (step S12). In FIG. 21, the edge amount is a multi-valued edge. If noise is not emphasized as an edge, a better image quality can be obtained. That is, the noise isolated point determination result (korits
If u) is true (YES in step S21), the edge amount is set to the specified value (step S23), otherwise, the edge amount remains unchanged (step S2).
2). With this correction means, the degree of freedom of the configuration of the combination of a plurality of processes can be increased.

In the image processing apparatus according to another embodiment of the present invention, when the target pixel is an isolated point, the data is not removed and replaced with an arbitrary smoothed value to make the noise isolated point unnoticeable and unnecessary. Try not to make a sharp edge part. Regarding the image processing apparatus according to the present embodiment, only the parts added or changed from the above embodiments will be described below.

FIG. 22 is a flowchart for explaining a judgment method according to another embodiment of the present invention. The data finally determined to be a noise isolated point is not set to 0, but is replaced with, for example, a smoothing filter result in this example in accordance with the image data. That is, in the determination of the noise isolated point, when the noise isolated point determination result (koritsu) is true (YES in step S31), the data is replaced with a smoothing filter result or the like (step S3).
3) If not, the data remains as it is (step S32).

Although various embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. For example, a program or a method for causing a computer to function as each unit in the above-described embodiments may be used. The present invention can be realized as a program to be executed, and further can be realized as a computer-readable recording medium on which the program is recorded.

[0046]

According to the first, second, seventh and eighth aspects of the present invention,
By referring to multi-value data as a parameter and making a decision,
The image data area surrounding the pixel of interest is not expanded by the resolution, and furthermore, noise isolated points are removed with high accuracy while the image data remains multi-valued. Good images can be obtained.

According to the third, seventh and eighth aspects, M × N
By determining by referring to the multi-valued data of the pixel as a parameter, character or picture data and isolated points can be identified with high accuracy without expanding the image data area surrounding the pixel of interest by the resolution. Good image quality without noise can be realized with a small memory.

According to the fourth, seventh and ninth aspects of the invention, L × K
By referring to and determining the multi-valued data of the pixel area as a parameter, the image data can be distinguished from the low-density halftone dot isolated points and the noise isolated points with high accuracy while remaining multi-valued.
In the subsequent processing, the multi-valued image data can be referred to as it is without requiring processing time or memory for calculating and storing the binarized data.

According to the fifth, seventh, and eighth aspects of the present invention, the determined result is reflected on the characteristic amount data of the target pixel, so that the noise isolated point is located downstream of the characteristic amount calculating means for the target pixel. Even in the case of removal, appropriate feature amount data can be passed to the subsequent processing, so that a flexible processing configuration can be adopted according to processing time, memory allocation, and the like.

According to the present invention, when the target pixel is an isolated point, the data is not removed,
By replacing it with an arbitrary smooth value, it is possible to make unnecessary edges unrecognizable while making noise isolated points inconspicuous, and it is possible to obtain a good image with no noticeable noise without affecting subsequent processing.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration example of an image processing unit of a digital copying machine according to an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a configuration example of an image processing unit of the digital copying machine according to the image processing apparatus according to the embodiment of the present invention.

FIG. 3 is a block diagram illustrating a configuration example of an image processing unit of the digital copying machine according to the image processing apparatus according to the embodiment of the present invention.

FIG. 4 is a block diagram illustrating a configuration example of an image processing unit of the digital copying machine according to the image processing apparatus according to the embodiment of the present invention.

FIG. 5 is a block diagram illustrating a configuration example of a noise removal circuit according to an embodiment of the present invention.

FIG. 6 is a diagram illustrating an example of a data area used for determination by a determination unit in FIG. 5;

FIG. 7 is a flowchart illustrating a determination method according to an embodiment of the present invention.

8 is a block diagram illustrating an example of an internal configuration of a first determination unit in FIG.

FIG. 9 is a block diagram illustrating another example of the internal configuration of the first determination unit in FIG. 5;

FIG. 10 is a diagram schematically illustrating an addition method of each line based on a configuration of a SIMD processor.

11 is a diagram in which FIG. 10 is imaged in two dimensions.

FIG. 12 is a diagram for explaining a determination method using data of A + 2, P, and B-2 shown in FIG. 11;

FIG. 13 is a diagram for explaining a determination method using the data of A, P, and B shown in FIG.

FIG. 14 is a diagram for explaining a determination method using data of A-2, P, and B + 2 shown in FIG. 11;

FIG. 15 is a diagram for explaining a method of determining after binarization according to a conventional technique.

FIG. 16 is a diagram for comparing and explaining a determination method according to the present invention with FIG.

FIG. 17 is a diagram illustrating an example of a region of a second determination unit.

18 is a diagram for explaining an operation of a second determination unit for making a determination in a second determination area of FIG.

FIG. 19 is a block diagram showing an example of a circuit for realizing the operation of FIG.

FIG. 20 is a flowchart for explaining a determination method according to another embodiment of the present invention.

FIG. 21 is a flowchart illustrating a determination method according to another embodiment of the present invention.

FIG. 22 is a flowchart for explaining a determination method according to another embodiment of the present invention.

[Explanation of symbols]

1,11,21,31 ... scanner, 2,12,22,3
2. Shading circuit, 3, 13, 23, 33, 46
... noise removal circuit, 4, 14, 24, 34 ... feature value extraction circuit, 5, 15, 25, 35 ... filter processing circuit, 4
7, 102: first determination means, 48, 103: second determination means, 49: determination result data, 51: area of a pixel of interest, 52: L × K pixel area, 61, 71: latest line,
62, 72: lines including the pixel of interest p, 63, 73: old lines, 64, 74: adders, 65, 75: level value determination circuits, 66, 76: reference density extraction circuits, 67, 77
... subtracters, 68, 78 ... comparators, 69 ... determination units, 79 ...
Judging section, 81, 102... Area of the first judging means, 82,
103: area of the second determination means, 91, 92, 93: pixel row, 101: image data, 104: inversion circuit, 105
... AND circuit, D1, D2... Lines, d1 to d6... Image data, F1 to F3.

Claims (8)

[Claims] 1. Each pixel data constituting image data is:
An image processing apparatus for determining whether an isolated point or a non-isolated point and removing pixel data determined to be an isolated point, wherein M × N pixels including target pixel data (M and N are integers) ), A first determining means for determining whether or not the target pixel is a candidate for an isolated point, and a region of L × K pixels (L and K are adjacent to the region of M × N pixels). Integer; K ≦
N) second determination means for correcting whether or not the pixel of interest is an isolated point by correcting the result of the first determination means from the multi-valued data of N), and a determination result by the second determination means And a data noise removing means for removing the pixel data of interest when the pixel data of interest is an isolated point in accordance with the following. 2. The image processing apparatus according to claim 1, further comprising a feature amount correcting unit configured to correct feature amount data corresponding to the target pixel data removed by the data noise removing unit.
The image processing apparatus according to any one of the preceding claims. 3. The method according to claim 1, wherein the first determining unit calculates a sum of data of adjacent t (t is an integer) pixels in the oldest line of the M × N pixels and the M × N pixels. A level value determining circuit for selecting a larger value as a level value from the sum of the data of the adjacent t pixels in the newest line, and the sum of the data of the adjacent t pixels including the target pixel. A reference density extraction circuit that extracts
When any or all of a vertical difference, a right diagonal difference, and a left diagonal difference between the extracted reference density value and the selected level value are greater than a predetermined threshold, 3. The image processing apparatus according to claim 1, wherein the target pixel is determined as a candidate for an isolated point. 4. The method according to claim 1, wherein the second determining unit calculates a continuous amount of pixels whose pixel data is smaller than a predetermined threshold value in the L × K pixel area adjacent to the M × N pixel area.
4. The method according to claim 1, wherein when the continuous amount is smaller than a predetermined value, the pixel of interest is determined as a non-isolated point even if the result of the first determination unit is an isolated point candidate. 2. The image processing device according to 1. 5. The feature amount correction unit replaces the feature amount data with an arbitrary value when the first determination unit and the second determination unit determine that the pixel of interest is an isolated point. The image processing apparatus according to claim 2, wherein: 6. When the target pixel is determined to be an isolated point based on the determination results of the first determination unit and the second determination unit, the data of the target pixel is replaced with an arbitrary smoothed value. The image processing apparatus according to claim 1, wherein: 7. A program for functioning as each means in the image processing apparatus according to claim 1. Description: 8. A computer-readable recording medium on which a program for functioning as each means in the image processing apparatus according to claim 1 is recorded.