JPH0973535A - Image processor and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute color emphasis for setting the fluctuation of emphasis effect owing to an object and the change of illumination light at the time of picking up an image to be small and for arbitrarily setting an emphasis color while the difference of the colors is clearly emphasized in a color image processing emphasizing the slight difference of the objective image. SOLUTION: A main component analysis part 2 analyzing a main component for the color density of the inputted image and deciding a threshold surface from an obtained result, an emphasis color setting part 4 setting a color emphasis vector E deciding the color after emphasis, a change chromaticity calculation part 3 calculating the change chromaticity of the respective image elements of image data by the threshold surface obtained by the main component analysis part 2 and the color emphasis vector E and a color emphasis part 5 emphasizing the colors of the respective image elements by using change chromaticity are provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and method for performing color enhancement on a specific target portion in a color image.

[0002]

2. Description of the Related Art Conventionally, it has been important to identify slight color differences in images in the inspection of surface defects by color images, medical image diagnosis and the like. For example, in the inspection of scratches on fruits and the like, the presence or absence of scratches is judged by utilizing the fact that the color of the scratches is different. In addition, in medical image diagnosis using an endoscope device or the like, doctors pay attention to subtle changes in color tone.

Therefore, many methods have been proposed so far in which color enhancement is performed in order to improve the visibility. For example, in color enhancement related to an endoscopic image processing device, Japanese Patent Laid-Open No. 2-11
4931, a color enhancement method for converting an RGB image into an HSV space and then performing histogram flattening;
A method of performing emphasis processing for each of the three elements of color, such as hue, saturation, and lightness, like 2-266028 is mentioned.

[0004]

However, the above method of flattening the histogram has a disadvantage that the image has an unnaturally different color tone compared to the input image. For example, if the histogram flattening process is applied to the saturation, the relatively low-saturation portion of the original image is emphasized with a lower saturation, so that the emphasized image becomes strangely whitish. However, it is inconvenient for a doctor to make such an extreme color change.

Another problem is that the effect varies depending on the change of the object and the illumination at the time of image pickup, and it is difficult to set the color of the emphasized part arbitrarily and variably. there were.

The present invention has been made in view of the above situation, and an object of the present invention is to clearly emphasize color differences in color image processing for emphasizing a slight color difference of a target image. However, it is still another object of the present invention to provide an image processing apparatus and method capable of reducing the variation of the enhancement effect due to the change of the object or the illumination lamp at the time of image pickup and performing the color enhancement in which the enhancement color can be arbitrarily set.

[0007]

Means for Solving the Problems That is, the present invention is as follows: (1) A principal component analysis unit for performing a principal component analysis on the color density of an input image containing at least a target portion; The determination means for determining the threshold plane from the result and the emphasis means for separating and enhancing the color density for the target portion in the image of the region divided by the threshold plane obtained by this determination means are provided. It is a thing.

In this way, the principal component analysis unit extracts the most representative color component in the image as the first principal component, and the deviation from the representative color component is the second principal component and the third principal component. It is represented by a component. Therefore, by setting a threshold plane in the color space using these principal component vectors, colors different from the average color can be separated and emphasized according to the image without setting a threshold value. Can be made.

(2) In the above item (1), the emphasizing means defines the distance from each pixel position forming the image in the feature space obtained by the principal component analysis to the threshold plane as the distance to the threshold plane. Distance calculation means for calculating the length of the perpendicular, vector input means for inputting an arbitrary transformation vector in the feature space, and coordinate transformation of the intersection coordinates of the curved surface of the perpendicular to the point position in the transformation vector direction. The coordinate conversion means is provided.

In this way, the difference from the average color can be emphasized to an arbitrary color by performing the emphasizing process according to the degree of color change using the color emphasizing vector that is arbitrarily input and set.

(3) In the above item (1), the emphasizing means is a specific region among regions divided by the threshold plane or a preset division plane in the feature space obtained by the principal component analysis. The color enhancement processing is applied only to the pixels belonging to.

In this way, only the arbitrary color enhancement is selected without performing unnecessary color enhancement processing on pixels in other regions which do not belong to the specific region divided by the threshold plane. You can

(4) In the above item (1), the density determination in which only the pixels whose density value in the input image meets the preset condition is processed as the preprocessing is executed, and A density value determination means for providing an image to the principal component analysis means or the enhancement means is further provided.

In this way, it is considered that the pixels suitable for color enhancement that meet the conditions preset by the density determination means in the preprocessing stage and the normal color of the object are not reflected. It is possible to realize the enhancement processing that is less affected by the lighting condition by excluding the pixels that do not reflect the normal color of the target object from the processing target. it can.

(5) In the above item (4), a process of providing the image to the subsequent stage and executing the process only when the pixel determined to be processed by the density value determining means meets a preset condition. The execution determination means is further provided.

In this way, by selecting only the pixels which meet the preset conditions by the processing execution determination means and providing them to the subsequent principal component analysis means or the emphasis means, the latter principal component analysis. The processing by the means or the emphasizing means can be made efficient.

(6) In the above item (5), the processing execution determining means selects the image data pixel only when the number of pixels to be processed reaches a certain ratio as compared with the total number of pixels of the image. It is provided at a later stage to execute processing.

With this configuration, when the number of pixels to be processed does not reach a certain ratio compared to the total number of pixels of the image, the supply of the pixel to the subsequent stage is stopped, so that it is necessary to emphasize. For a very small number of pixels determined not to exist, the processing in the subsequent stage can be omitted, and the processing in the subsequent principal component analysis means or the emphasis means can be made more efficient.

(7) Principal component analysis processing for performing principal component analysis on the color density of the input image containing at least the target portion, and determination processing for determining the threshold plane from the results obtained by this principal component analysis processing. And an emphasizing process for separating and emphasizing the color density with respect to a target portion in the image in the region divided by the threshold plane obtained by the determining process.

In this way, the most representative color component in the image is extracted as the first principal component by the principal component analysis processing, and the deviation from the representative color component is the second principal component and the third principal component. It is represented by a component. Therefore, by setting the threshold plane in the color space using these principal component vectors, the color different from the average color can be separated and emphasized according to the image without setting the threshold value. can do.

(8) In the above item (7), the emphasizing process is defined as the distance from each pixel position forming the image in the feature space obtained by the principal component analysis to the threshold plane, to the threshold plane. A distance calculation process for calculating the length of a perpendicular line, a vector input process for inputting an arbitrary transformation vector in the feature space, and coordinate conversion of intersection coordinates of the perpendicular line with the curved surface to a point position in the transformation vector direction. It has a coordinate conversion process.

In this way, the difference from the average color can be emphasized to an arbitrary color by performing the emphasizing process according to the degree of color change by using the color emphasizing vector arbitrarily input and set.

(9) In the above item (7), the density determination in which only the pixels whose density value in the input image meets the preset condition is processed as the preprocessing is executed, and The image is further provided with a density value determination process for providing the principal component analysis process or the enhancement process. By doing this, it is possible to know the pixels that are likely to be color-emphasized and that meet the conditions preset by the density determination processing as preprocessing.

[0024]

BEST MODE FOR CARRYING OUT THE INVENTION

(First Embodiment) A first embodiment of the present invention will now be described with reference to the drawings.
An embodiment will be described. FIG. 1 shows the circuit configuration of the apparatus. This apparatus includes an image input unit 1 for inputting image data from another image pickup apparatus (not shown), and an R for input image data.
A principal component analysis unit 2 that performs a principal component analysis in the GB space and determines a threshold plane using the obtained principal components, and image data is configured using the threshold plane obtained from the principal component analysis unit 2. A color change degree calculation unit 3 that calculates a deviation of the pixel data from an average color, an emphasis color setting unit 4 that sets a color emphasis vector that determines the color after emphasis, and a color emphasis unit that performs color emphasis according to the degree of color change. 5, image output unit 6 for outputting color-enhanced image data, image input unit 1, principal component analysis unit 2, color change degree calculation unit 3, color enhancement unit 5, and image output unit 6 by bus connection to image data Has a data bus 7 for transferring data, a controller 8 for overall control of other circuits, and an image memory 9 connected to the data bus 7 and capable of storing image data for a plurality of frames under the control of the controller 8. .

Next, the operation based on the above configuration will be described. FIG. 2 mainly shows the flow of the entire processing operation controlled by the control unit 8, and at the beginning thereof, image data from another image pickup device (not shown) is first input to the image input unit 1 (step A1). ).

Here, the image pickup device for obtaining the input image is not particularly limited, and may be a television camera, an infrared camera or the like, or a medical image pickup device such as an ultrasonic diagnostic device, an MRI device or an endoscope device. May be In the image input unit 1,
The input image data is A / D-converted for each pixel and digitized, and then stored as original image data in the image memory 9 via the data bus 7.

The principal component analysis unit 2 then performs principal component analysis on the original image data stored in the image memory 9 (step A2). The principal component analysis performed by the principal component analysis unit 2 is a method of extracting the most representative component of the target information, and the detailed processing contents are as shown in FIG.

That is, the principal component analysis unit 2 first calculates a covariance matrix of RGB density values of each pixel of image data (step B1), and then calculates an eigenvalue and an eigenvector of the calculated covariance matrix (step B2). ). The covariance matrix of RGB density values is

[0029]

[Equation 1] And E (·) in this equation represents an average value. The eigenvalues and eigenvectors of the obtained covariance matrix are λi and υi that satisfy the following equation (2), respectively.

[0030]

[Equation 2]

The above covariance matrix is 3 × as shown in equation (1).
Since it is a matrix of 3, there are three types of eigenvalues and eigenvectors obtained. It is known that the eigenvalues of the covariance matrix are all positive real numbers, of which the eigenvector corresponding to the largest eigenvalue is the first principal component. Similarly, the eigenvector corresponding to the second largest eigenvalue is the second principal component, and the eigenvector corresponding to the smallest eigenvalue is the third principal component.

The meaning of each principal component varies depending on the type of image data, but when the object is the inner wall of the digestive tract, for example, as in a digestive endoscopy image, the direction of the first principal component is the image. It represents the average color of the whole data, and the closer it is to the origin, the darker it is. And
The second principal component and the third principal component often represent a deviation from an average color.

The advantage of performing the principal component analysis is that even if the color of the object changes or the illumination at the time of imaging changes,
The most representative component can be extracted, and the influence of the change of the color of the object and the change of the illumination at the time of image pickup is less than that in the case where the color enhancement is performed using the absolute value.

When the calculation of the eigenvalues and eigenvectors of the covariance matrix is completed, the threshold plane is determined using the calculated principal components (step B3). In this embodiment,
As shown in FIG. 5, a plane whose normal vector is the second principal component is a threshold surface. By determining the threshold plane in this way, the deviation from the average color can be separated and emphasized. This threshold surface equation defines the second principal component as (υ _R , υ
_G , υ _B ) is expressed by the following equation (3).

[0035]

(Equation 3)

Depending on the purpose of processing and the nature of the image, a plane having the first principal component or the third principal component as a normal vector may be used as the threshold surface. For example, when a plane in which the first principal component is the normal vector is determined as the threshold surface, brightness enhancement can be performed.

Further, it is not always necessary to use a flat surface as the threshold surface, and a curved surface such as a parabolic surface or a hyperboloid may be used. However, after the threshold plane is determined in step A2, the emphasis color setting unit 4 sets the color emphasis vector E for determining the color after emphasis (step A3). The color emphasis vector E may be set only once before the processing is started, but the set value can be changed for each image.

Next, the color change degree calculating section 3 calculates the color change degree of each pixel of the image data (step A4), and the color enhancing section 5 performs color enhancement using the calculated color change degree (step A).
5). FIG. 4 shows the contents of the processing by the color change degree calculating unit 3 and the color emphasizing unit 5 in steps A4 and A5. First, as shown in FIG. 6, the color change degree calculating unit 3 converts the RGB values corresponding to the pixel values. From the points in space, calculate the coordinate position corresponding to the "foot" of the perpendicular line that has been dropped on the determined threshold surface (step C1), and calculate the distance between these two points, that is, the length of the perpendicular line. This is defined as the discoloration degree d (step C2). That is, the degree of color change d is expressed by the following equation (4).

[0039]

(Equation 4) The color emphasis part 5 is defined by using the discoloration degree d.
Color enhances.

In the color emphasizing process in this embodiment, as shown in FIG. 7, a vector V ₀ formed by a vertical line drawn from the point of each pixel in the RGB space to the threshold plane is emphasized according to its size. This is performed by converting to the direction of the color emphasis vector E set by the color setting unit 4 (step C
3).

[0041] That is, the point _{(V R, V G, V} B) the position coordinates of the perpendicular foot dropping off from the threshold plane (p _R, p _G,
p _B ), the emphasizing process is as in the following expression (5), and the density values (V ′ _R , V ′ _G , V ′ _B ) after emphasizing are obtained. That is,

[0042]

(Equation 5) (However, k: enhancement coefficient. When k is increased, the image is more strongly emphasized.) Each pixel after conversion obtained by the color conversion process for color emphasis in this way is sequentially formed as the image data after emphasis. It is stored in the image memory 9.

Processing up to the above color conversion (steps C1 to C1)
The image data enhancement process is performed by repeatedly performing C3) until it is determined that the process has been completed for all pixels (step C4).

Here, it is possible to perform emphasis processing only on pixels belonging to one of the areas divided by the threshold plane in the RGB space. That is, as shown in FIG. 8, it is also possible to perform emphasis processing only on pixels belonging to either one of the regions 1 and 2 divided with the threshold plane S as a boundary.

This is because, for example, when the RGB space is divided into a reddish color region and a light blue color region by the determined threshold surface, the pixels belonging to the former reddish color region. Only the colors are emphasized, and the pixels belonging to the light blue color area are not emphasized. For example, by setting the presence / absence of the color emphasis for each area, it is unnecessary. Color enhancement is avoided, processing is made more efficient, and an unnatural enhanced image is prevented.

In this case, for the pixels belonging to the non-emphasized area, the original image data stored in the image memory 9 is directly transferred to the emphasized image data stored in the image memory 9. Is. Further, instead of the emphasizing process of the above formula (5), the following formula (6), that is,

[0047]

(Equation 6) By performing the emphasis process defined by, the color close to the average color is not emphasized so much, while the color greatly different from the average color is emphasized more strongly. Furthermore, more generally, the following equation (7), that is,

[0048]

(Equation 7) Various enhancement effects can also be obtained by applying the enhancement processing defined by and making f (d, k) an appropriate function.

After the color enhancement processing in step A5 as described above is completed, the image output unit 6 reads the enhanced image data from the image memory 9 and converts it into an analog value by D / A conversion. Output as (Step A
6) The above is the end of this series of processing.

As described above, in the first embodiment, the threshold value corresponding to the input image is determined for the input image, and the emphasized color is arbitrarily set to obtain the average color. It is possible to realize color enhancement processing that effectively enhances the difference.

(Second Embodiment) A second embodiment of the present invention will be described below with reference to the drawings. FIG. 9 shows the circuit configuration thereof, which is basically the same as that shown in FIG. 1 described above, and therefore, the same reference numerals are given to the same portions and the description thereof will be omitted.

In addition to the image input unit 1, the principal component analysis unit 2, the color change degree calculation unit 3, the color enhancement unit 5, the image output unit 6, the image memory 9, and the density value determination unit, the data bus 7 is connected to the data bus 7. 1
0 is connected to send and receive image data. A control unit 8'is provided to integrally control each of these circuits.

Next, the operation based on the above configuration will be described. FIG. 10 mainly shows the flow of the entire processing operation controlled by the control unit 8 ', and at the beginning, image data from another image pickup device (not shown) is input to the image input unit 1 (step D1).

Here, the image pickup device for obtaining the input image is not particularly limited, and may be a television camera, an infrared camera or the like, or a medical image pickup device such as an ultrasonic diagnostic device, an MRI device or an endoscope device. May be In the image input unit 1,
The input image data is A / D-converted for each pixel and digitized, and then stored as original image data in the image memory 9 via the data bus 7.

For the original image data stored in the image memory 9, the density value determination unit 10 next makes a density value determination (step D2). The density value determination performed by the density value determination unit 10 is to determine whether or not the density value of the pixel in the image data is a value close to the maximum value or more. If the pixel density value is greater than or equal to the maximum value, it may be saturated when the image data is captured because the illumination is too close to the object or due to reflected light. It is considered that it does not reflect the normal color of the object. Therefore, by excluding such pixels from the processing target, it is possible to realize the enhancement processing that is less affected by the illumination state.

The specific processing of this density value determination is executed as shown in FIG. That is, first the RGB of that pixel
Of the respective concentration values of "Vr_ma"
x, Vg_max, Vb_max ”(step E1), and only the confirmed pixels are set as the pixels to be emphasized (step E).
The process 2) is repeatedly executed until it is determined that all the pixels forming the image data have been processed (step E3). Analysis is performed (step D3).

The principal component analysis performed by the principal component analysis unit 2 is a method of extracting the most representative component of the target information, and the detailed processing contents are the same as in FIG. In this case, the principal component analysis unit 2 executes the principal component analysis processing only for the pixels set as the processing target in the density determination processing, and the covariance matrix is expressed by the following equation instead of the equation (1).

[0058]

(Equation 8) Is defined by

However, after the threshold plane is determined in step D3, the emphasis color setting unit 4 sets the color emphasis vector E for determining the color after emphasis (step D4).
The color emphasis vector E may be set only once before the processing is started, but the set value can be changed for each image.

Next, the color change degree calculating section 3 calculates the color change degree of each pixel of the image data (step D5), and the color enhancing section 5 performs color enhancement using the calculated color change degree (step D).
6). FIG. 12 shows the contents of the processing by the color change degree calculation unit 3 and the color emphasis unit 5 in steps D5 and D6. First, the color change degree calculation unit 3 confirms that the pixel is a processing target ( Step F1), as shown in FIG. 6, from the points in the RGB space corresponding to the pixel values, the coordinate position corresponding to the "foot" of the perpendicular drawn to the determined threshold surface is calculated (Step F2 ), The distance between these two points, that is, the length of the perpendicular line is calculated and defined as the discoloration degree d (step F3). The degree of color change d is defined by the above equation (4), and the color emphasis unit 5 performs color emphasis using this degree of color change d.

The color emphasizing process in this embodiment is the same as that shown in FIG.
In the direction of the color emphasis vector E set in the emphasis color setting unit 4, the vector V ₀ formed by the perpendicular line drawn from the point of each pixel in the RGB space to the threshold plane as shown in FIG. The conversion is performed (step F4).

That is, the coordinates of the position of the foot of the perpendicular line drawn from the point (V _R , V _G , V _B ) to the threshold plane are (p _R , p _G ,
p _B ), the emphasizing process is as shown in the above equation (5), and the density values (V ′ _R , V ′ _G , V ′ _B ) after emphasizing are obtained.

The converted pixels obtained by the color conversion process for color enhancement in this way are sequentially stored in the image memory 9 as the constituents of the image data after enhancement. It should be noted that the processing of steps F2 to F4 is not performed for the pixels determined not to be processed in step F1 and the density value of the original image data is retained and set as the density value of the emphasized word. Is.

The processing up to the above color conversion (steps F1 to F1)
The image data enhancement process is performed by repeatedly performing F4) until it is determined that all pixels have been processed (step F5).

Here, it is possible to perform emphasis processing only on pixels belonging to one of the areas divided by the threshold plane in the RGB space. That is, as shown in FIG. 8, it is also possible to perform emphasis processing only on pixels that belong to either one of the regions 1 and 2 which are divided with the threshold surface S as a boundary.

This is because, for example, when the RGB space is divided into a reddish color region and a light blue color region by the determined threshold surface, the pixels belonging to the former reddish color region. Only the colors are emphasized, and the pixels belonging to the light blue color area are not emphasized. For example, by setting the presence / absence of the color emphasis for each area, it is unnecessary. Color enhancement is avoided, processing is made more efficient, and an unnatural enhanced image is prevented.

In this case, for the pixels belonging to the non-emphasized area, the original image data stored in the image memory 9 is directly transferred to the emphasized image data stored in the same image memory 9. Is.

Further, instead of the emphasizing process of the equation (5),
By applying the emphasis processing defined by the above equation (6),
Colors close to the average color are not emphasized so much, while colors greatly different from the average color are emphasized more strongly.

Further, various emphasizing effects can be obtained by applying emphasizing processing more generally defined by the above equation (7) to make f (d, k) an appropriate function. Then, after the color enhancement processing in step D6 as described above is finished, the image output unit 6 reads the enhanced image data from the image memory 9, converts it into an analog value by D / A conversion, and outputs it as an image signal. (Step D7), the series of processing is completed.

Further, instead of the processing of FIG. 10, FIG.
, The density value determination in step D2 and step D3
During the processing of the principal component analysis of, the total number of pixels set as the processing target in the image data is set to a threshold value T
If it is determined that the number of pixels to be processed is too small for the entire image by performing the process execution determination as to whether or not h is exceeded (step D11), the process starts from step D3.
The original image data may be output as it is without performing the enhancement processing of D6.

As described above, in the second embodiment, the density value determination unit 10 sets the conditions for the pixels to be subjected to the principal component analysis, whereby the adverse effects of the bad imaging conditions can be eliminated, and further processing is performed. By performing the execution determination, the enhancement processing is performed when it is determined that the target object is not normally imaged because the number of pixels to be processed is too small, such as when the illumination is too close to the entire target portion during imaging. It is possible to prevent it.

Although the RGB color space is used as the feature space in the first and second embodiments, other color spaces such as HSV space, L * u * v * space, and xyz space are used. Good. In addition, the present invention can be variously modified without departing from the scope of the invention.

[0073]

As described above in detail, according to the present invention, in the color image processing for emphasizing the slight color difference of the target image, while clearly emphasizing the color difference, it is possible to illuminate the target object or the illumination at the time of imaging. It is possible to provide an image processing apparatus and method capable of reducing the variation of the enhancement effect due to the change of the lamp and performing the color enhancement capable of arbitrarily setting the enhancement color.

[Brief description of drawings]

FIG. 1 is a block diagram showing a circuit configuration according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing overall processing according to the same embodiment.

FIG. 3 is a flowchart showing a subroutine of principal component analysis processing of FIG.

FIG. 4 is a flowchart showing a subroutine of color change degree calculation and color enhancement processing of FIG.

FIG. 5 is a diagram explaining an operation according to the embodiment.

FIG. 6 is a diagram illustrating an operation according to the embodiment.

FIG. 7 is a diagram explaining an operation according to the embodiment.

FIG. 8 is a diagram illustrating an operation according to the embodiment.

FIG. 9 is a block diagram showing a circuit configuration according to a second embodiment of the present invention.

FIG. 10 is a flowchart showing overall processing according to the same embodiment.

11 is a flowchart showing a subroutine of the density value determination processing of FIG.

12 is a flowchart showing a color change degree calculation and color enhancement subroutine of FIG.

FIG. 13 is a flowchart showing another overall process according to the embodiment.

[Explanation of symbols]

 1 ... Image input unit 2 ... Principal component analysis unit 3 ... Discoloration degree calculation unit 4 ... Enhancement color setting unit 5 ... Color enhancement unit 6 ... Image output unit 7 ... Data bus 8, 8 '... Control unit 9 ... Image memory 10 ... Concentration value determination unit

Claims (9)

[Claims] 1. A principal component analysis means for performing a principal component analysis on the color density of an input image including at least a target portion, and a determination means for determining a threshold plane from a result obtained by the principal component analysis means. An image processing apparatus, comprising: an emphasizing unit for separating and emphasizing color density with respect to a target portion in the image of the region divided by the threshold plane obtained by the determining unit. 2. The emphasizing means calculates a length of a perpendicular to the threshold plane as a distance from each pixel position forming the image in the feature space obtained by the principal component analysis to the threshold plane. A calculation means; a vector input means for inputting an arbitrary transformation vector in the feature space; and a coordinate transformation means for transforming the coordinates of the intersection of the perpendicular line with the curved surface into a point position in the transformation vector direction. The image processing apparatus according to claim 1, wherein: 3. The emphasizing means applies color only to pixels belonging to a specific area among areas divided by the threshold plane or a preset division plane in the feature space obtained by the principal component analysis. The image processing apparatus according to claim 1, wherein the image processing apparatus performs enhancement processing. 4. The density determination in which only the pixels whose density value in the input image meets a preset condition is processed as preprocessing, and the image after execution is subjected to the main component analysis means or The image processing apparatus according to claim 1, further comprising density value determination means provided to the enhancement means. 5. The method further comprises a process execution determining unit that provides the image to a subsequent stage to execute the process only when the pixel determined to be processed by the density value determining unit meets a preset condition. The image processing apparatus according to claim 4, wherein 6. The processing execution determination means provides the image data pixel to the subsequent stage to execute the processing only when the number of pixels to be processed reaches a certain ratio as compared with the total number of pixels of the image. The image processing device according to claim 5, wherein 7. A principal component analysis process for performing a principal component analysis on the color density of an input image including at least a target portion, and a determination process for determining a threshold plane from the result obtained by this principal component analysis process. An image processing method comprising: an emphasizing process for separating and emphasizing color density for a target portion in the image of the region divided by the threshold surface obtained by the determining process. 8. The emphasis processing is a distance for calculating a length of a perpendicular to the threshold plane as a distance from each pixel position forming the image in the feature space obtained by the principal component analysis to the threshold plane. It has a calculation process, a vector input process for inputting an arbitrary transformation vector in the feature space, and a coordinate transformation process for transforming the coordinates of the intersection of the perpendicular line with the curved surface to a point position in the transformation vector direction. The image processing method according to claim 7, wherein 9. A density determination targeting only pixels whose density values in the input image meet a preset condition is performed as preprocessing, and the image after execution is subjected to the principal component analysis processing or The image processing method according to claim 7, further comprising a density value determination process provided for the enhancement process.