JPH09149246A - Picture processing method and picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent the deterioration of picture quality by changing the resolution of an original at the time of separating an image area for the first time and the resolution of an original at the time of separating the image area after second time. SOLUTION: In pre-scanning, an edge separation processing and a dot separation processing are executed in parallel on a thinned picture. At the time of scanning, the color picture is reproduced by executing scanning for four times in one drum type and for once in four drum type, Whether a character is included in the original or not is judged in pre-scanning. When the character is included, thinning processing circuits 8 and 9 are turned off. An edge separation circuit 11, a dot separation circuit 12 and a color separation circuit 10 operate in parallel on picture data which is not thinned and respective judgment results are inputted to a judgment circuit 13. The judgment circuit 13 judges a black character, a color character and a pattern by means of the presence or absence of the edge, the dot and the color of a notice element. Then, a processing is switched for respective picture elements in accordance with the result. Thus, the deterioration of picture quality can be prevented.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and apparatus using prescan.

[0002]

2. Description of the Related Art At present, in an image processing apparatus such as a color copying machine, when processing and reproducing an original in which a character area and a picture area are mixed, the area of the image is recognized by a so-called image area separation technique. However, the processing is switched between the character portion and the pattern portion in one document.

As a conventional color copying machine using such an image area separating technique, for example, there is one described in JP-A-3-63886. This image area determination method for color copiers that does not involve prescan determines whether or not there is a sharp density change in the vicinity of the pixel of interest, and there are consecutive sharp density change points in a specific direction. The edge area is extracted by determining whether or not
Further, it is a method of determining a halftone dot area by a combination of density changes in a specific direction.

As an example of controlling the image area separation during the main scan based on the prescan information, for example, Japanese Patent Application Laid-Open No. 4-1
There is an image processing method described in Japanese Patent No. 80350. In this method, it is determined whether or not a halftone image is included in the prescan, and if it is determined that the halftone image is not included, the halftone determination in the main scan is turned off. As a result, it is possible to reduce the deterioration of the image quality due to the local erroneous halftone dot determination.

[0005]

However, the above-mentioned color copying machine has the following problems depending on the image area separation algorithm and the input resolution. That is, the first problem is that erroneous separation occurs in a halftone dot pattern having a low number of lines such as a newspaper, and image quality is greatly deteriorated. The second problem is that the step edge in the photographic printing paper original is erroneously separated as a character, resulting in a large deterioration in image quality.

The main reason for the above first problem is as follows. The halftone dot area determination method described in the above publication detects halftone dot pixels as a halftone dot pixel when the density in the vicinity of the target pixel changes, and performs correction based on the density information. On the other hand, although it is a density change peculiar to a halftone dot, there are many similar portions in the character portion. High density of dots 175
Characters and halftone dots can be separated with high accuracy by correction based on density information, as long as they are lines.
When the number of lines becomes low, the density of the halftone dot pixels in the halftone dot part and the similar halftone dot pixels in the character part (that is, the wrong halftone dot pixels) becomes similar, and as a result, the character part and halftone dot The point part is erroneously separated, and the above-mentioned first problem occurs.

The main reason for the second problem is as follows. In the character edge determination unit means described in the above publication, the step edge in the printing paper is determined as a character from the feature amount. That is, in the 3 × 3 mask, since the determination is made based on the difference value between the pixel of interest and its neighboring pixels, the character portion and the edge in the photographic paper pattern cannot be separated by the threshold processing.

By the way, if it is known in advance that characters will not be mixed in a document for which accurate separation cannot be expected in the current image area separation processing, for example, by turning off the image area separation processing, The above-mentioned problem does not occur at least for a manuscript having only a picture. The present invention has been made by paying attention to this point, and an object of the present invention is to carry out image area separation at least twice, and to make a halftone dot original or photographic paper original having a particularly low number of lines by the first image area separation judgment. An object of the present invention is to provide an image processing method and apparatus that prevent image quality deterioration by determining whether or not a character is included and feeding back the result to the determination of image area separation in the subsequent stage.

Another object of the present invention is to carry out image area separation at least twice and utilize a plurality of image area separation judgments to prevent image quality deterioration especially for a halftone dot original having a low number of lines. And to provide a device.

[0010]

In order to achieve the above object, according to the invention of claim 1, an original is read out as a digital signal and image area separation is performed on the original a plurality of times.
In the image processing method for switching the processing within the original based on the separation result, the resolution of the original when the image area separation is performed first and the resolution of the original when the image area separation is performed for the second time and thereafter are changed. It is characterized by that.

According to the second aspect of the invention, the original is read out as a digital signal, the original is subjected to image area separation a plurality of times, and the second and subsequent image area separation is controlled based on the first separation result. In the image processing method for switching the processing within the document based on the separation result after the second time, the resolution of the document when first performing the image area separation and the document resolution when performing the image area separation after the second time It is characterized by changing the resolution.

According to the third aspect of the invention, the original is read out as a digital signal, the original is subjected to image area separation a plurality of times, the first separation result is held, and the held separation result is used to perform 2 In the image processing method of performing the image area separation after the second time and switching the processing in the original based on the separation result after the second time, the resolution of the original when the image area is first separated and the second and subsequent times. It is characterized by changing the resolution of the original when the image area separation is performed.

According to a fourth aspect of the present invention, the resolution of the original when the image area separation is first performed is coarser than the resolution of the second and subsequent times.

According to a fifth aspect of the invention, the resolution is changed by performing optical scaling in the sub-scanning direction.

According to the sixth aspect of the present invention, the original is read out as a digital signal, the original is subjected to a plurality of image area separations, and the second and subsequent image area separations are controlled based on the first separation result. In an image processing apparatus that switches processing within the original document based on the separation result after the second time, first means for performing the image area separation, and second means for performing the image area separation after the second time. It is characterized by having.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be specifically described below with reference to the drawings. <Embodiment 1> FIG. 1 shows the structure of Embodiment 1 of the present invention.
In the figure, 1 is an image input device such as a scanner, and 2 is a Log that converts a linear reflectance signal into a linear density signal.
A conversion circuit, a filter circuit including a smoothing filter and an edge emphasis filter, a color correction circuit for converting a cmy signal into YMC of a complementary color, a UCR circuit for performing processing for subtracting a K signal from each color material signal, Reference numeral 6 denotes a dither circuit for expressing halftones using dither for characters and pictures, and reference numeral 7 denotes an image output device such as a printer.

Reference numerals 8 and 9 are thinning processing circuits for generating an image in which pixels are thinned out every other pixel in the main scanning direction and the sub-scanning direction of the image, 10 is a color separating circuit for separating a color region from the original, and 11 is an original. An edge separation circuit that separates character edges from the inside, 12 a halftone dot separation circuit that separates halftone dots from the document, 13 a determination circuit that determines an area based on the separation result of the separation circuit, and 14 counts character edge pixels Counting circuit, 1
Reference numeral 5 is a determination circuit for determining whether or not a document contains characters based on the count value, and 16 is a determination result storage circuit for storing the determination result of the determination circuit 15.

First, the outline of FIG. 1 will be described. The image input device 1 (resolution is 400 dpi) such as a scanner outputs rgb data (8 bits each) which is almost linear with respect to the reflectance. Next, Lo is pre-processed for performing color correction at a later stage.
The logarithmic conversion is performed in the g conversion circuit 2, and the density linear c
Generate my data.

The filter circuit 3 is composed of a circuit in which a smoothing filter and an edge enhancement filter as shown in FIGS. 5 (a) and 5 (b) are connected in series. , If it is a color character, the smoothing filter is made to be through and only the edge enhancement processing is performed.

The color correction circuit 4 converts the cmy signal into a complementary color YMC signal in consideration of the characteristics of the image input device 1 and the image output device 7. As a color correction method, linear approximation, so-called masking method, tetrahedral interpolation method, triangular prism interpolation method, and the like have been proposed, and color correction is performed using these methods. At this time, the K signal is simultaneously calculated from min (C, M, Y).

The UCR circuit 5 switches the processing as follows based on the area determination signal from the determination circuit 13. That is, for the black character area, K is made to be through and C = M
= Y = 0. For the color character area, K is through, C is through, M is through, and Y is through. For other than the above, that is, K '= 0.6 × K C' (M ', Y') = C (M, Y) ) -K '.

Then, the dither circuit 6 prepares two kinds of dither tables (dither matrices) having different sizes (not shown), and when the signal from the judging circuit 13 is a black character or a color character, the size is changed. Of the dither table (eg, 1 × 1), and in other cases, the dither table of larger size (eg, 2 × 2) is used to send the data to the image output device 7.

The image area separating section, which is a main characteristic part of the present invention, will be described in detail below. The image area separation unit according to the first embodiment includes an edge separation circuit 11, a halftone dot separation circuit 12, and a color separation circuit 10. Among them, the edge separation circuit 11 and the halftone dot separation circuit 12 are described in, for example, the paper Image area separation method for mixed images of patterns (dots and photographs) "(Journal of the Institute of Electronics, Information and Communication Engineers Vo
l. J75-DI1 No. 1 pp39-47 1992
The method of "4.2 Edge region detection" described in January, 2002) is used.

The thinning-out processing circuits 8 and 9 perform thinning-out of an image as a pre-processing of the edge separation circuit 11 and the halftone dot separation circuit 12 every other pixel here in the main and sub-lines. The reason is,
In the conventional image area separation technology, when an image is input by a scanner of about 400 dpi, a halftone dot (generally 100 lines or less) having a low frequency is erroneously determined as a character area. This is because the dot density of halftone dots and the density of erroneous halftone dots having a data structure locally similar to that of halftone dots in the character area are substantially the same.

Therefore, the thinning-out process is performed, and the pseudo 20
An image equivalent to 0 dpi is created, and separation performance equivalent to the separation accuracy for halftone dots of 100 lines or more before thinning is obtained. On the other hand, by thinning out, the separation accuracy of the character edge is slightly lowered in the fineness of the separation boundary (at the level where edge separation is performed in a 2 × 2 block), but the important thing here is that the manuscript contains characters. This is to have a sufficient and sufficient separation accuracy for determining whether or not there is such a condition, which is satisfied in the present invention.

The edge separation circuit 11 is a circuit for detecting character edges in a document. In this embodiment, for example, the method described in the above-mentioned paper is used as the character edge detection method.

According to this method, edge enhancement is applied to input image data of 64 gradations, and then ternarization is performed by using two types of fixed threshold values, and 3
The continuity of the black pixels and the white pixels after the binarization is detected by pattern matching, and when one or more of the black continuous pixels and the white continuous pixels are present in the block of 5 × 5 pixels, the target block is defined as the edge region. If not, it is determined as a non-edge region.

The halftone dot separation circuit 12 is a circuit for detecting a halftone dot (print) area in the original. This detection method also uses the “4.1 halftone dot area detection” method described in the above-mentioned paper. This method focuses on the fact that the density change between the color halftone dot area and the black and white halftone dot area is significantly different from the density change in the character area, and performs peak pixel detection, halftone dot area detection, and halftone dot area correction. This is to separate the dot areas.

The peak pixel is detected by, for example, in a block of 3 × 3 pixels, the density level L of the central pixel is higher or lower than that of all the surrounding pixels, and
The density levels a and b of the diagonal pixels existing diagonally across L and the central pixel are | 2 × L−a−b |> TH for all four pairs.
When it is (fixed threshold), the central pixel is set as the peak pixel. The detection of the halftone dot area is performed, for example, in four blocks each including 4 × 4 pixels, if there are two or more blocks including the peak pixel, the target block is set as the halftone dot candidate area, and the other blocks are non-halftone dots. Determined as a candidate area. After determining the halftone dot / non-halftone dot candidate region, if four or more of the nine blocks centering on the target block are halftone dot candidate regions, the target block is set as the halftone dot region, and otherwise the target block is selected. It is a non-dot area.

The same effect can be obtained by using the halftone dot area determination using the DCT performed with a block size of about 8 × 8, which does not depart from the gist of the present invention.

The color separation circuit 10 is a circuit for detecting a chromatic color portion in a document. As the color separation method, for example, the chromatic / achromatic color area separation method described in JP-A-3-260878 is used. In this method, the difference between the R, G, and B digital signals is obtained, and when the maximum difference is equal to or larger than a predetermined threshold value, it is determined that the color is chromatic.

The operation of the first embodiment will be described below. (During prescan) In the configuration of the embodiment of FIG. 1, the processing unit that operates during prescan is surrounded by a dotted line. The prescan referred to here is not limited to the case where the color scanner actually scans the document, but also includes the case where the image is stored in the memory and the image is used a plurality of times, that is, the first use.

As shown by the dotted line in FIG. 1, the edge separation process and the halftone dot separation process are performed in parallel on the thinned image, that is, on the image in which both the main and sub pixels are thinned out every other pixel. Next, in the determination circuit 13, when the pixel of interest is an edge and is not a halftone dot, it is determined as a true character edge pixel. Then, the same determination is made for all the images, and the counting circuit 14 counts the character edge pixels. When the count value is equal to or larger than a predetermined threshold value, the determination circuit 15 determines that the original includes characters, and holds the result in the determination result storage circuit 16 until the main scan.

Regarding the thinning-out in the sub-scanning direction, optical scaling (for example, 50%) may be performed instead of the above-described digital signal thinning-out. As a result, the processing time during prescan is shortened.

(At the time of main scanning) The 1-drum type color copying machine scans a total of 4 times for each CMYK plate.
The four-drum type performs one scan to reproduce a color image. However, as described above, when all images are stored in the memory and used multiple times, the scanner does not operate.

By the above-described prescan, it is determined whether or not the original contains characters. If a character is included, normal image area separation is performed. That is, the thinning processing circuits 8 and 9 are turned off, and the edge separation circuit 11 for the image data that is not thinned, the halftone dot separation circuit 12 for the image data that is not thinned, and the color separation circuit 10 It operates in parallel and each judgment result is input to the judgment circuit 13.

In the judging circuit 13, when the pixel of interest is an edge and is not a halftone dot but an achromatic color, it is judged as a black character, and when it is an edge and not a halftone dot but a chromatic color, It is determined to be a color character, and the others are determined to be a pattern. As described above, the processing is switched for each pixel according to the result.

On the other hand, when the character is not included, for example, as an example of the control, the image area separation processing is turned off and the entire surface is subjected to the pattern processing. Alternatively, it is also an effective method to reset the parameters that are easy to determine the pattern as the respective parameters of the image area separation before starting the main scan. For example, the parameter TH for halftone dot separation (pp. 42 of the above-mentioned paper) is reduced. Alternatively, the ternarization value of edge separation (pp. 44 of the same paper) is set to a value that is difficult to determine as a character, for example, white pixels are set to 16 to 8 and black pixels are set to 32 to 48. This reduces the possibility of erroneous separation in the pattern area. However, although it tends to cause erroneous separation with respect to characters, it does not substantially cause a problem because the characters do not exist, or the existence of the characters is considerably small.

As described above, in the first embodiment, the halftone dot original having a low number of lines does not particularly contain characters, or even if it contains characters, the image quality is significantly deteriorated with respect to the original having a small area. Will be reduced.

<Second Embodiment> FIG. 2 shows the structure of a second embodiment of the present invention. Image reproduction system (that is,
Since the steps from the image input device to the image output device) are the same as those in the first embodiment, the description thereof will be omitted.

The image area separation unit of the second embodiment includes an edge separation circuit 11, a halftone dot separation circuit 12, and a color separation circuit 10. These edge separation circuit 11, halftone dot separation circuit 12, and color separation circuit 10 are implemented. It has the same function as in Example 1. The storage device 17 provided according to the second embodiment activates the halftone dot separation circuit 12 during prescanning and stores the separation result for each pixel.
The storage device 17 has a memory capacity of A4 size document (29
7 mm × 210 mm), when reading 16 pixels per 1 mm (equivalent to 400 dpi) and thinning out every other pixel in the pre-scan, as a result of halftone dot separation,
A memory capacity of (297 × 8) × (210 × 8) = 3991680 bits is required.

The operation of the second embodiment will be described below. (During prescan) The processing unit that operates during prescan is surrounded by a dotted line. As shown in the dotted line in FIG. 2, the halftone dot separation circuit 12 operates for the thinned image, that is, for the image in which both the main and sub pixels are thinned out every other pixel. Then, the storage device 17 stores the halftone dot result until the main scan. As a result, as described in the first embodiment, the halftone dot area having a low number of lines is accurately separated.

As in the first embodiment, optical scaling (for example, 50%) may be performed for thinning in the sub-scanning direction. As a result, the processing time during prescan is shortened.

(During Main Scan) At the time of main scan, the thinning processing circuit 9 is turned off, and the halftone dot separation circuit 12 operates for image data that is not thinned. As described above, the storage device 17 stores the information on the halftone dot area having a low number of lines. The stored information is ORed with the result of the halftone dot separation performed on the image data that is not thinned out. That is, the former highly accurately separates halftone dots with a low number of lines (approximately 50 to 100 lines) from the image data, and the latter highly accurately separates halftone dots with a high number of lines (approximately 100 to 200 lines). Then, by taking the OR with the OR circuit 18,
Both will make up for it.

The determination circuit 19 determines that the target pixel is an edge, not a halftone dot, and an achromatic color, that it is a black character,
If it is an edge and is not a halftone dot but a chromatic color, it is determined to be a color character, and the others are determined to be a pattern. As described above, the processing is switched for each pixel according to the result.

As described above, in the second embodiment, regardless of the number of lines,
It is possible to significantly reduce the image quality deterioration due to the erroneous separation of the halftone dot original.

<Third Embodiment> FIG. 3 shows the structure of a third embodiment of the present invention. The image reproducing system (that is, from the image input device to the image output device) is the same as that in the first embodiment, and the description is omitted.

The image area separation unit of the third embodiment includes an edge separation circuit 11, a halftone dot separation circuit 12, a color separation circuit 10, and a white background separation circuit 20, of which the edge separation circuit 11 and the halftone dot separation circuit 12 are included. The color separation circuit 10 has the same function as that of the first embodiment. In this embodiment, the types of image area separation used during prescan and main scan, that is, the feature amount is changed.

The white background separation circuit 20 provided in the image area separation section of this embodiment is a circuit for detecting an area surrounded by a white background. Here, for example, the method described in Japanese Patent Laid-Open No. 4-248766 is used. In this method, a group of white pixels in the vicinity of the pixel of interest is detected by pattern matching, and when matching, the pixel of interest is made the active pixel. This is a method in which the number of active pixels is counted in a block centered on the pixel of interest, and if there is even one active pixel in the block, the entire block is determined as a white background area. What is important as the function of white background separation used here is that step edges in a pattern are not erroneously recognized as characters.

The operation of the third embodiment will be described below. (During prescan) The processing unit that operates during prescan is surrounded by a dotted line. During the prescan, the edge separation circuit 11 and the white background separation circuit 20 operate in parallel. This white background separation has the following features. FIG. 4 (a),
(B), (c) is a figure explaining the character / non-character determination using white background separation.

FIG. 14A is the same diagram as that shown in FIG. 14 of the above-mentioned publication, and is a diagram in which the pixel of interest is determined to be a character. That is, the white background separation circuit 20 detects a cluster of white pixels at a position separated from the target pixel by L pixels to the left and right, and when the target pixel is detected as a character candidate region by the edge separation circuit 11, the target pixel is detected. Is determined to be a character area.
The character determined in this way is a thin character.

On the other hand, the bold character edge in (b) is
In the case of the figure, since a cluster of white pixels is not detected at a position distant from the pixel of interest by L pixels to the right, it is not detected as a character edge. As described above, a bold character edge is not detected, but if a thin character exists in another portion, there is no particular problem in determining whether or not a character is included. Also, if only bold characters are present, it is determined to be non-character, so
In the present embodiment, the entire pattern is processed, but if the characters are thick, there is no particular problem in terms of readability.

On the other hand, with respect to the photographic paper original, as shown in (c), the step edge does not detect a cluster of white pixels at a position distant from the pixel of interest by L pixels to the right, so that it is not erroneously detected as a character. . In other words, it is important and necessary to prevent erroneous detection of character edges, especially step edges in the printing paper, even if some of the characters cannot be detected.

When the pixel of interest is an edge and there is a white background, the judging circuit 21 judges as a true character edge pixel. Then, the determination is performed on all the images, and the character edge pixels are counted by the counting circuit 22. When the count value is equal to or larger than the predetermined threshold value, the determination circuit 23 determines that the original includes characters, and holds the determination result in the determination result storage unit 24 until the main scan.

(During Main Scan) It is possible to know whether or not the original contains characters by the above prescan. When a character is included, it functions as a normal image area separation circuit. That is, for the image data after Log conversion,
The edge separation circuit 11, the halftone dot separation circuit 12, and the color separation circuit 10 operate in parallel, and the respective judgment results are input to the judgment circuit 21. Here, the white background separation circuit 20 does not operate during the main scan. The reason is that there is a possibility that characters of various thicknesses (sizes) may be included in the manuscript, so if the characters are judged to be edge / non-edge depending on the character thickness and processed, This is because a large deterioration in image quality occurs.

The determination circuit 21 determines that the pixel of interest is a black character when the pixel of interest is an edge and is not a halftone dot and an achromatic color, and it is a color character when it is an edge and not a halftone dot and a chromatic color. And other than them are determined to be patterns. Depending on the result, the processing is switched for each pixel as described above.

When no character is included, for example, as an example of the control, the image area separation is turned off, and the entire surface is patterned. Alternatively, it is also effective to reset each parameter of the image area separation before starting the main scan so that it is easy to determine the pattern.

As described above, according to the present embodiment, the deterioration of the image quality is significantly reduced when the photographic printing paper original does not include the character or the area including the character is a small area.

[0059]

As described above, claims 1 and 2,
According to the invention described in 4, the character frequency in the original is recognized during the pre-scan, and the image area separation during the main scan is controlled according to the result, so that the character is not included (or the small area is included). It is possible to significantly improve the image quality of a halftone printed document.

According to the first, third, and fourth aspects of the present invention, a halftone dot area having a low number of lines is recognized during the prescan, a halftone dot area having a high number of lines is recognized during the main scan, and the logical sum of the two is calculated. Therefore, the halftone dot area can be separated with high accuracy regardless of the number of lines, and the image quality can be improved.

According to the invention described in claim 5, the processing time at the time of prescan is shortened.

According to the sixth aspect of the invention, the character frequency in the original is recognized during the prescan, and the image area separation during the main scan is controlled according to the result, so that the character is not included (or Small area even if included)
The image quality can be significantly improved.

[Brief description of the drawings]

FIG. 1 shows a configuration of a first exemplary embodiment of the present invention.

FIG. 2 shows a configuration of a second embodiment of the present invention.

FIG. 3 shows a configuration of a third embodiment of the present invention.

4 (a), (b), and (c) are diagrams illustrating determination of character / non-character using white background separation.

FIG. 5 shows an example of a filter circuit, in which (a) is a smoothing filter and (b) is an edge enhancement filter.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image input device 2 Log conversion circuit 3 Filter circuit 4 Color correction circuit 5 UCR circuit 6 Dither circuit 7 Image output device 8 and 9 Thinning processing circuit 10 Color separation circuit 11 Edge separation circuit 12 Halftone dot separation circuit 13 and 15 Judgment circuit 14 Counting circuit 16 Judgment result storage circuit

Claims (6)

[Claims] 1. An image processing method in which an original is read out as a digital signal, image areas are separated from the original a plurality of times, and processing within the original is switched based on the result of the separation. Resolution of the original when performing
An image processing method characterized by changing the resolution of an original when separating the image areas after the first time. 2. An original document is read out as a digital signal, image area separation is performed on the original document a plurality of times, image area separation for the second and subsequent times is controlled based on the first separation result, and separation for the second and subsequent times is performed. In the image processing method for switching the processing within the original based on the result, it is possible to change the resolution of the original when the image area separation is first performed and the resolution of the original when the image area separation is performed for the second time and thereafter. Characterized image processing method. 3. An original is read out as a digital signal, image area separation is performed on the original a plurality of times, the first separation result is held, and the second and subsequent image area separation is performed using the held separation result. In the image processing method of switching the processing in the original based on the separation result of the second time and thereafter, the resolution of the original when the image area separation is performed first and the image area separation after the second time An image processing method characterized by changing the resolution of the original document. 4. The image processing method according to claim 1, wherein the resolution of the original at the time of performing the image area separation is coarser than the resolution at the second and subsequent times. 5. The image processing method according to claim 1, wherein the resolution is changed by performing optical scaling in the sub-scanning direction. 6. An original is read out as a digital signal, image area separation is performed on the original a plurality of times, image area separation for the second time and thereafter is controlled based on a result of the first separation, and separation for the second time and thereafter is performed. An image processing apparatus that switches processing within the original based on the result includes first means for performing image area separation first and second means for performing image area separation after the second time. Image processing device.