JP2006031688A - Image processing device, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a pictorial drawing image formed by processing image data captured from a photograph and expressing the image data in a natural touch. <P>SOLUTION: As to pixels included in the image data, the respective pixels are grouped by using color and the like as one of determination criteria, and the image data are divided into a plurality of basic areas. For each of the basic areas, its characteristics including characteristics such as an average color and a position of the gravity center are calculated, and based on the characteristics, a plurality of basic areas are combined into an integrated area. Characteristics of each of the integrated areas are calculated, and a plurality of basic areas inside the integrated area, for example, are combined into a plurality of combined areas. Alternatively, the integrated area is partially divided for forming a partial area, and the combined area or the partial area can be selected as a unit area of expression in addition to the basic area and the integrated area. For each of the unit area of expression decided according to the selected result, a color of pixel included in the unit area is decided based on the characteristics of the unit area of expression. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus, an image processing method, and an image processing program for performing predetermined processing, for example, painting-like processing, on image data.

  2. Description of the Related Art Conventionally, various image processing processes such as a color conversion process that changes a pixel color and a distortion conversion process that distorts an image are performed on a read image data by a scanner. For example, from the input image data and the result of dividing the image data into regions, the pixels are divided into groups for each divided region, group features are calculated and recorded for each group, and the recorded group features are processed. An image processing / synthesizing device that automatically assigns a new and rich expression to image data using a group of pixels corresponding to the stroke of a painting as a unit of conversion by constructing a new pixel from the processed group feature Has been developed (see Patent Document 1). In this case, features are grouped into groups, features are processed in units of groups, and pixels are assigned to groups to store the processed features. The original image such as a photographic image can be processed into various brush-touch images.

JP-A-10-11569 (page 3-9-, FIG. 1-14)

  However, in the image processing / synthesizing apparatus described above, since a group of pixels corresponding to the stroke of a painting is used as a unit of conversion, it is possible to create a style similar to a point drawing drawn with a large stroke or a small stroke. There wasn't. In addition, the touch created by constructing a new pixel from the group feature has also become relatively simple.

  An object of the present invention is to provide an image processing apparatus, an image processing method, and an image processing program capable of appropriately giving natural brush strokes of various sizes and shapes to image data based on image information. It is.

  In the present invention, for example, image data obtained by taking a photograph is grouped by grouping together pixels that are close in color and distance, and thereby divided into a plurality of basic regions. Next, basic regions that are similar to each other in the plurality of basic regions are further grouped into a plurality of integrated regions. As the simplest method of the present invention, roughly speaking, in the integrated region thus obtained, it is predetermined whether the integrated region itself is used as a unit of expression or the basic region is used as a unit of expression. The processing method (corresponding to “processing instruction preset” in the embodiment) or the processing method manually determined by the operator. For example, for a part of the background, the integrated region is used as a unit of expression. For a part of a person, a basic area is expressed as a unit of expression (the color of a pixel included in the area is determined according to the characteristics of the area).

  A further feature of the present invention is how to create an integrated area. The integrated area is further processed to combine, for example, a plurality of basic areas in the integrated area into a plurality of combined areas. Alternatively, the integrated area is partially divided to form a partial area, and in addition to the basic area and the integrated area, a combined area or a partial area can be selected as a unit area for expression, thereby ensuring greater freedom of expression. In this way, images with various styles of paintings and natural brushstrokes can be obtained.

A basic aspect of the present invention is a basic area division processing means for grouping each pixel with respect to pixels included in the image data using at least a color as one of the determination criteria, thereby dividing the image data into a plurality of basic areas. When,
For each divided basic region, basic region feature calculating means for calculating features of the basic region including at least color features based on pixel data included in the basic region;
Based on at least the color of the basic region features calculated by the basic region feature calculating means, region integrating means for integrating a plurality of basic regions to produce an integrated region;
An integrated region feature calculating means for calculating a feature of the integrated region including at least a color feature for each integrated region;
An image processing area instruction unit that determines whether the integrated area is a unit area of expression or a basic area included in the integrated area is a unit area of expression according to the integrated area;
For each unit region of the expression determined based on the selected result, an expression unit that determines the color of the pixel included in the unit region based on the feature of the unit region of the expression is provided.
An image processing apparatus according to the present invention includes a divided image data storage unit that stores divided image data obtained as a result of processing of a basic region division processing unit that divides image data into a plurality of basic regions;
Basic area feature calculating means for calculating the characteristics of the basic area for each basic area based on the divided image data stored in the divided image data storage means;
Basic area feature storage means for storing the basic area feature calculated by the basic area feature calculating means for each basic area;
Area integration means for producing an integrated area by integrating a plurality of basic areas based on the basic area features stored in the basic area feature storage means;
From an integrated region feature storage unit that stores an integrated region feature calculated by the integrated region feature calculation unit from the integrated region integrated by the region integration unit and the divided image data stored in the divided image data storage unit. Image feature extraction means comprising:
An integrated area design map storage means for storing an integrated area design map written out when calculating an integrated area feature by the integrated area feature calculating means;
Stored in the integrated region feature storage means using a calculation method for changing a calculation formula or a coefficient included in the calculation formula in accordance with information of the integrated region design map stored in the integrated region design map storage means. Calculated from the integrated region features of the integrated region and the basic region features stored in the basic region feature storage means, or in addition to these, the new region features created by combining or dividing these region features Expression means for performing image processing based on the result.

  In the image processing device according to the present invention, the image data is divided into a plurality of basic areas, the basic area features of the basic area are calculated for each basic area, and the basic areas are further calculated based on the calculated basic area features. The integrated area feature is calculated from the integrated basic area, and the integrated area design map is also calculated based on the information obtained in the integration process. In addition to the integrated region features of the integrated region stored in the integrated region feature storage unit and the basic region features stored in the basic region feature storage unit using a calculation method for changing the included coefficient, Since the processed image data is calculated based on the result calculated from the new area feature created by combining or dividing these area features, it is the basic from the basic area feature. Can be used to create large-size touches from integrated region features, intermediate-size brushes from combined region features, and small-size touches from segmented region features. . And in the integrated area design map, information to be used as reference for storing which part of the image should be drawn with which brush size is stored. Objects drawn by drawing with a brush of a small size that can be described in detail, and by drawing with a brush of various sizes according to the expression in a part that is effective for changing the impression of appearance by adding various styles of painting This makes it possible to create works that are expressive while retaining the characteristics of

  The expression means described above calculates the characteristics of the pixels such as color information constituting the image, and the two area features of the integrated area feature and the basic area feature described above, or one of the area features and the image feature. The position information is used as an input value, and a predetermined calculation 1 is performed. As a result, in addition to the basic area feature of the basic area, the integrated area feature of the integrated area, or both, the maximum or minimum value is given. It is desirable to calculate the pixel feature using all or some of the values calculated from the difference between the minimum value and the second smallest value as input values.

  In this case, the pixel whose difference between the maximum value and the second largest value is 0 is a pixel that may have the characteristics of either region, that is, the position of the pixel is the border of the brush stroke, and the pixel that is not 0 is the value Can be regarded as an index of how far away from the boundary, so that different expressions can be given according to the distance from the periphery of the brush stroke.

  The above-described image feature extraction unit includes a region combining unit that selects and combines two or more basic regions from a plurality of basic regions integrated into the same integrated region, and a combined region combined by the region combining unit. A combined region calculation unit that calculates a combined region feature from the divided image data stored in the image data storage unit; and a combined region feature storage unit that stores the combined region feature calculated by the combined region calculation unit. It is desirable that

  In this case, in addition to the integrated region feature that becomes a large brush stroke and the basic region feature that becomes a small brush stroke, a combined region feature that can create a brush with an intermediate size becomes available.

  The image feature extraction means described above includes, in addition to each means described in the image feature extraction means, a partial area dividing means for dividing the integrated area, the basic area, or the combined area into a plurality of partial areas, respectively. In addition to the integrated region feature stored in the region feature storage unit, the basic region feature stored in the basic region feature storage unit, or the combined region feature stored in the combined region feature storage unit, A partial region feature calculating unit that calculates a partial region feature from each partial region with reference to the divided image data stored in the divided image data storage unit, and a partial region feature storage unit that stores the calculated partial region feature It is desirable to have.

  In this case, since the three types of region features are further divided, it is possible to express with a finer touch than the basic region that has been the finest so far. The divided region feature also serves as basic information for expressing a single region as a set of a plurality of partial regions.

  The expression means described above is stored in the integrated area feature of the integrated area stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, and the combined area feature storage means. 3 types of region features, and, in addition to them, 4 types of region features from among the partial region features stored in the partial region feature storage means, or any of several types thereof The basic region which gave the minimum value as a result of performing a predetermined calculation 2 that is considered to indicate the minimum value when the most preferable region feature is selected, using the region feature and pixel position information as input values The basic area feature, or the integrated area feature of the integrated area, or the combined area feature of the combined area, or in addition to the minimum value of the calculation result of the predetermined calculation 2 and the second All values are calculated from the difference between the values, or it is preferable to calculate the pixel characteristics of the pixels constituting the image as input some of them.

  In this case, by calculating the pixel features from the combined region features and partial region features in addition to the integrated region features and basic region features, it becomes possible to create small brush strokes and medium size brush strokes. Power increases.

The image feature extraction unit described above is stored in the integrated region feature of the integrated region stored in the integrated region feature storage unit, the basic region feature stored in the basic region feature storage unit, and the combined region feature storage unit. Types of combined region features and three types of region features, or in addition to these, four types of region features calculated by the partial region dividing means, or any number of them Each region feature of
A feature changing means for changing each area feature based on a value calculated from each area feature of the basic area, the integrated area, the combined area, or the partial area located around the basic area, the integrated area, the combined area, or the partial area. It is desirable to have

  In this case, since one brush is created from one area feature, it is necessary to obtain a painting-like area feature in order to create a painting-like touch, but one area feature is a feature of the area surrounding the area. Since it changes depending on the value, for example, as a region feature, the direction of the region is influenced by the direction of the surrounding region, and when it changes, a wavy brush is created.

  The above-mentioned area integration means finds a basic area without a mark and puts a mark 1 on it. Mark a basic area where the difference in feature quantity between the basic areas adjacent to the basic area is smaller than the first threshold value, and further change the feature quantity between the basic areas adjacent to the marked basic area. Repeat the process of marking 1 even for basic areas where the difference is less than the first threshold. Finally, the basic areas marked with mark 1 are combined into a temporary integrated area 1. Find a basic area that is not marked, add a mark 2 to the basic area, repeat the same process, and finally repeat until there are no unmarked basic areas. Next, paying attention to a temporary integration area in which the number of basic areas included in the temporary integration area is larger than the first number, a difference in feature amount between adjacent basic areas is larger than the first threshold, and the second threshold A smaller basic area is collected by the same method, and the new temporary integrated area that has been collected includes the basic area that was included in the previous temporary integrated area, and the basic area that has been added to the new temporary integrated area. When the value obtained by dividing the number of areas by the number of basic areas gathered in the temporary integrated area of interest is larger than 1 and smaller than the specified value 1, it is desirable to formally determine the temporary integrated area as an integrated area. .

  In this case, when the integrated area is not determined at once, the integrated area is not changed significantly even if the integration process is performed in the second integration process that is performed by changing the threshold value for the first time without setting the integrated area once. Because it is determined as a formal integration area, the gradation part is rarely cut off in the middle, and only the integration area with high accuracy can be extracted, so that areas that are considered to be homogeneous as paintings can be combined as an integration area become.

  After the second temporary integration area has been determined, the area integration unit described above uses the second threshold value as a new first threshold to determine the third temporary integration area and the second temporary integration area Is officially determined as an integrated area, and after that, the same operation is repeated several times, and when the process is repeated, only the remaining basic areas excluding the basic area determined as the integrated area in the previous process are integrated. For example, when the determination process as the nth formal integration area is completed, a serial number is assigned to the integration area, and the number and the nth uniform integration area are recorded in the integration area design map. Is desirable.

  In this case, since the integration is repeated while increasing the threshold value, it is possible to integrate regions that could not be integrated only twice. In addition, since the region integrated with a small number of repetitions is a more uniform region, if this region is recorded in the design map, the information is added to the pixel. It can be used effectively.

  The area integration means described above focuses on a temporary integration area whose number of basic areas is larger than the first number and smaller than the second number after the n-th repetitive processing, and is located near the boundary of the temporary integration area. The absolute value of the average difference between the basic area features of the basic area within the temporary integration area of interest and the basic area features of the basic area adjacent to and outside the temporary integration area of interest within the area is specified. When the value is larger than 2, further assign a serial number to the temporary integration area, record the serial number and the nth comparison integration area on the integration area design map, and select the temporary integration area from the target of the integration process. It is preferable to perform the (n + 1) th repeated processing after the exclusion.

  In this case, when considering the design, the integrated area with a small area and strong contrast with the surrounding area is conspicuous and becomes an important area, so such areas are integrated and determined and recorded in the integrated area design map By doing so, both the integrated area and the integrated area design map can be used for expression.

  The area integration means described above excludes basic areas having basic area characteristics larger than the specified value 3 in advance before performing all the processing, and the information on the excluded basic areas is added to the previous integrated area design map. It is desirable to describe it.

  In this case, the basic region riding on the edge part between the two objects should not be integrated into either, but to achieve this, the basic region riding on the edge part must be extremely elongated. If this feature is expressed as a basic region feature, if this feature is larger than the specified value 3, it can be excluded from the candidates without integration, and this information Since the integrated area design map is retained, it can be used for expression.

  The area combining means described above can combine the number of basic areas to be combined as a single combined area with those basic areas combined with a plurality of adjacent basic areas grouped into the same integrated area by the area integrating means. It is preferable that the relationship with the basic area feature is changed according to information described in the integrated area design map of the integrated area to which the basic area belongs.

  In this case, whether or not to combine the basic areas is the part that corresponds to the thickness of the brush drawn in terms of expression, so it should be decided with reference to the information that the image itself has, but the information is integrated area Since the design map has it, you can use it to select a brush with an appropriate thickness according to the style.

  The expression means described above is stored in the integrated area feature of the integrated area stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, and the combined area feature storage means. In addition to these three types of region features, or in addition to these, four types of region features from among the partial region features calculated by the partial region dividing means, or any of several types of each of them A region feature is selected based on information described in the integrated region design map of the integrated region in which each region is included, and from the region features selected based on the information described in the integrated region design map. It is preferable to calculate the pixel feature by switching the calculation formula for calculating the pixel feature.

  In this case, since the integrated area design map has information on the edge part and important parts even in a small area, avoid such a part by using the information of the integrated area design map and use various image styles. By giving an expression, it is possible to create an expression that does not impair the content of the original picture.

  First, since it is important from which type of region the pixel feature is calculated, the pixel feature is calculated from the basic region feature or from the integrated region feature of the integrated region that includes the basic region. By determining whether to calculate pixel features from the combined region where the basic region is combined based on the information of the integrated region design map, it is possible to draw finely where to draw Become.

  Furthermore, since the style can be changed by selecting a calculation formula that calculates pixel features from the selected features, it is possible to select an appropriate calculation formula based on the information in the integrated region design map. Expression will be added.

  The partial area dividing means described above divides the integrated area, the basic area, or the combined area into a plurality of partial areas, and integrates the integrated areas stored in the integrated area feature storage means for each partial area. When calculating a partial region feature with reference to a region feature, or a basic region feature stored in the basic region feature storage unit, or a combined region feature stored in the combined region feature storage unit, A different calculation formula is selected according to the information described in the integrated region design map of the integrated region to which it belongs, and the partial region feature of the partial region is calculated by the partial region calculation means using it, and the calculated partial region The features are preferably stored in the partial area feature storage means.

  In this case, the partial area is created separately from the basic integrated area, basic area, and combined area, but how to create this partial area is also an important element that changes expression. By giving an appropriate calculation formula according to the described information, an appropriate expression can be given as a result.

  The feature changing means described above is stored in the integrated area feature of the integrated area stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, and the combined area feature storage means. A plurality of mathematical formulas for changing each region feature selected from four types of region features or any number of types among the combined region features and the partial region features calculated by the partial region dividing means In accordance with the information described in the integrated area design map of the integrated area stored in the integrated area feature storage means, an appropriate mathematical expression is selected, and each feature amount is calculated using the selected mathematical expression. It is preferable to change.

  In this case, by making the information described in the integrated area design map available to the feature changing means for further changing the color and outline of each area feature that has already been calculated, it is possible to determine which feature of the image. It will become clear whether there should be a change.

  For each of the integrated region feature, combined region feature, basic region feature, integrated region, combined region, and basic region that is a description unit of each region feature, the region serial number and the statistics of the pixels included in the region are obtained. It is preferable to include as an element, at a minimum, the average color, the barycenter of the pixels included in each area, and the number of pixels included in each area.

  In this case, various statistics can be considered as necessary for the expression of paintings, but if they are not included in any area, there are elements that greatly reduce their expressive power. , Avoid the risk of loss of expressiveness.

  The expression means described above may perform expression processing on image data by using the integrated area feature stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, or the combination. A data input means for dividing the image data into a plurality of areas so that different processing can be applied by the operator's instruction, with the image range calculated from the combined area feature stored in the area feature storage means as a unit. It is preferable to provide.

  In this case, it is possible to directly indicate an image range to be expressed to the image processing apparatus in units of the integrated area, basic area, and combined area.

  The expression means described above is stored in the integrated area feature of the integrated area stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, and the combined area feature storage means. The shape of the brush created from each of the three area features can be changed variously using the partial area feature calculated from each of the three area features by the partial area dividing means. Is preferred.

  In this case, the complexity of the shape of the brush can be changed according to the number of partial areas created from one area feature. For example, if three partial areas are created, the use of the partial area features enables the deformation of the brush according to the quadratic curve.

  The expression means described above is stored in the integrated area feature of the integrated area stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, and the combined area feature storage means. Three types of region features, or in addition to them, four types of region features from among the partial region features calculated by the partial region dividing means, or any of several types of regions among them Using the feature and image position information as input values, a predetermined calculation 3 is performed that minimizes the value when the most preferable feature is selected, n is a natural number, and the result of the calculation is the first to nth smallest value N basic region features of n basic regions, n integrated region features of n integrated regions, n connected region features of n connected regions, or some of them, and All or some of the differences between the smallest value of the previous calculation result and the second smallest value (n−1 values calculated from the difference between the smallest value and the (n−1) th smallest value) It is desirable to calculate a pixel feature using as an input value.

  In this case, when determining one pixel, the pixel feature can be calculated in consideration of not only one region but also features of a plurality of regions considered to have an influence on the pixel. The expression can be realized.

An image processing method according to the present invention is an image processing method for processing an image using a computer,
A divided image data storage step in which the computer stores divided image data obtained as a result of the processing of the basic area dividing processing step of dividing the image data into a plurality of basic areas;
A basic region feature calculating step for calculating a feature for each basic region based on the divided image data stored in the divided image data storage step;
A basic region feature storing step in which the computer stores the basic region feature calculated in the basic region feature calculating step for each region;
An area integration step in which the computer creates an integrated area by integrating a plurality of basic areas based on the basic area features stored in the basic area feature storage step;
An integrated region feature storing step in which the computer stores the integrated region integrated in the region integrating step;
Based on the result calculated from the two area features of the integrated area feature stored in the integrated area feature storage step and the basic area feature stored in the basic area feature storage step, the image data is converted into the image data. And an expression step for performing expression processing on the image.

An image processing method according to the present invention is an image processing method for processing an image using a computer,
A divided image data storage step in which the computer stores divided image data obtained as a result of the processing of the basic area dividing processing step of dividing the image data into a plurality of basic areas;
A basic region feature calculating step for calculating a feature for each basic region based on the divided image data stored in the divided image data storage step;
A basic region feature storing step in which the computer stores the feature calculated in the basic region feature calculating step for each basic region;
The computer integrates a plurality of basic areas based on the basic area features stored in the basic area feature storage step to create an integrated area, and an area integration step of creating an integrated area design map;
An integrated region storage step in which the computer stores the integrated region integrated in the region integration step;
A region combining step in which the computer combines a plurality of basic regions based on the integrated region feature stored in the integrated region feature storing step to create a combined region;
A combined region feature storing step in which the computer stores the combined regions combined in the region combining step;
The integrated region feature stored in the integrated region feature storing step, the basic region feature stored in the basic region feature storing step, and the combined region feature stored in the combined region storing step for storing the combined region, or any arbitrary A partial region dividing step of adaptively dividing the plurality of regions into a plurality of portions while referring to the integrated region design map, and calculating a partial region feature for each partial region;
A feature changing step in which the computer changes the integrated region feature, or the basic region feature, or the combined region feature, or the partial region feature; and the computer, the integrated region feature, or the basic region feature, or An expression step for adaptively expressing the combined area feature or the partial area feature with reference to the integrated area design map.

An image processing program according to the present invention is an image processing program for processing an image,
Divided image data storage means for storing divided image data obtained as a result of processing of a basic area dividing processing means for dividing the image data into a plurality of basic areas;
Basic area feature calculating means for calculating a feature for each basic area based on the divided image data stored in the divided image data storage means;
Basic area feature storage means for storing, for each basic area, the basic area feature calculated by the computer in the basic area feature calculating means;
Area integration means for creating an integrated area by integrating a plurality of basic areas based on the basic area features stored in the basic area feature storage means by the computer;
An integrated area feature storage means for storing an integrated area integrated by the computer in the area integration means;
Based on the result calculated from the two area features, the integrated area feature of the integrated area stored in the integrated area feature storage means and the basic area feature stored in the basic area feature storage means, the image data is converted into the image data. On the other hand, the computer is caused to function as a means for performing expression processing.

An image processing program according to the present invention is an image processing program for processing an image,
Divided image data storage means for storing divided image data obtained as a result of processing of a basic area dividing processing means for dividing the image data into a plurality of basic areas;
Basic area feature calculating means for calculating a feature for each basic area based on the divided image data stored in the divided image data storage means;
Basic region feature feature storage means for storing, for each basic region, the features calculated by the computer in the basic region feature calculation means;
The computer integrates a plurality of basic areas based on the basic area features stored in the basic area feature storage means to create an integrated area, and an area integration means for producing an integrated area design map;
An integrated area feature storage means for storing an integrated area integrated by the computer in the area integration means;
Area combining means for creating a combined area by combining a plurality of basic areas based on the integrated area features stored in the integrated area feature storage means by the computer;
A combined region feature storage means for storing a combined region combined by the computer in the region combining means;
All of the integrated region feature stored in the integrated region feature storage unit and the basic region feature stored in the basic region feature storage unit and the combined region feature stored in the combined region feature storage unit that stores the combined region, or A partial region dividing means for adaptively dividing a plurality of arbitrary regions into a plurality of portions while referring to the integrated region design map, and calculating a partial region feature for each partial region;
Feature change means for the computer to change the integrated region feature, or the basic region feature, or the combined region feature, or the partial region feature;
Means for adaptively performing image processing on the integrated region feature, the basic region feature, the combined region feature, or the partial region feature while the computer refers to the integrated region design map. It is what makes a computer function.

Furthermore, the present invention is used in a computer, and in a program for processing image data,
For each pixel included in the image data, grouping each pixel using at least a color as one of the determination criteria, thereby dividing the image data into a plurality of basic regions;
Calculating a feature of the basic region including at least a color feature for each divided basic region;
For the divided basic regions, a basic region group in which features of the basic regions that are adjacent to each other and are adjacent to each other with respect to the one basic region of interest are smaller than the first threshold is detected, and the basic regions are detected. Similarly, a group of temporary integrated areas is grouped, and a group of basic areas that are also grouped with respect to a basic area other than the one basic area is detected and set as a temporary integrated area. And getting the steps
The first threshold value is made larger than that in the step of obtaining the first temporary integration region, the same steps as the above step are performed, and a plurality of temporary integration regions are obtained, and these are used as the second temporary integration region. The first temporary integration area and the second temporary integration area include a common basic area, and the second temporary integration area corresponds to the number of basic areas included in the first temporary integration area. When the rate of increase in the number of basic areas is smaller than a preset value, determining the provisional integrated area as a formal integrated area;
After the second temporary integration area is determined, the first threshold value is further increased and the third temporary integration area is determined in the same manner. Similarly, the second temporary integration area is formalized. After that, when the same operation is repeated several times, the integration process is continued only for the remaining basic areas excluding the basic area determined as the integrated area in the previous process. Steps,
After the process of determining as a formal integration area is completed, for each formal integration area, recording the number of stages indicating the number of temporary integration processes determined from the temporary integration area to the formal integration area;
Calculating a feature of the integrated region including at least a color feature based on pixel data included in the basic region for each officially determined integrated region.

In this program, according to the integration area, determining whether the integration area is a unit area for expression or a basic area included in the integration area is a unit area for expression;
For each unit region of the expression determined based on the selected result, a step of determining a color of a pixel included in the unit region based on the feature of the unit region of the expression may be included. The determination of whether the integrated area is the unit area for expression or the basic area included in the integrated area is the unit area for expression is performed, for example, according to the number of stages of the integrated area.

  According to the present invention, for image data, pixels that are close in color and distance are grouped together and divided into a plurality of basic areas, and the basic areas that are similar to each other in the plurality of basic areas are grouped together. Are grouped into multiple integrated areas, and whether the integrated area itself is used as the unit of expression or the basic area is used as the unit of expression is specified automatically or manually. Can be realized.

  Also, the image data is divided into a plurality of basic areas, integrated areas, combined areas, and partial areas, and the characteristics of each are calculated, information on the integrated area design map, information input from the data input means, or predetermined Because the pixel features that make up the image are calculated after deciding which pixel feature calculation formula to use for the pixels that make up the image based on any of the feature quantities in each region according to the formula , It becomes possible to include various expressions in one work.

  Hereinafter, an image processing apparatus according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an image processing apparatus according to an embodiment of the present invention. The image processing apparatus shown in FIG. 1 is composed of a computer having a ROM (Read Only Memory), a CPU (Central Processing Unit), a RAM (Random Access Memory), an external storage device, an input device, a display device, a communication device and the like. ing. In this computer, an image processing program stored in advance in an external storage device or the like is executed by a CPU or the like, whereby an image data input unit 1, an image feature extraction unit 2, a basic region division unit 3, and a basic region feature calculation unit 4. , Basic region feature storage unit 5, region integration unit 6, integrated region feature calculation unit 7, integrated region feature storage unit 8, region combination unit 9, combined region feature calculation unit 10, combined region feature storage unit 11, partial region division unit 12, partial region feature calculation unit 13, partial region feature storage unit 14, expression unit 15, feature change unit 16, integrated region design map storage unit 17, image processing region instruction unit 18, image processing region preset storage unit 19, data input Functions as a unit 20, an image data display unit 21, an image data output unit 22, a divided image data storage unit 23, and a processing procedure storage unit 24.

  The image data input unit 1 includes an input device such as a mouse, a keyboard, and a tablet, and a communication device such as a communication interface. The image data input unit 1 generates image data such as a scanner and a digital camera in accordance with a user command using the input device. Accept as an image, input the original image to the basic region dividing unit 3 constituting the image feature extraction unit 2, and directly accept the processing procedure describing the image processing procedure and the parameters required for the processing from the keyboard, The processing procedure is read from the hard disk, and the processing procedure is stored in the processing procedure storage unit 24 including a RAM or an external storage device.

  Here, various image data can be used as the original image, and image data expressed in RGB color system, XYZ color system, L * a * b * color system can be used, or one or a plurality of image data can be used. A sheet of image data may be used.

  The basic area dividing unit 3 divides the input original image into a plurality of basic areas by a known area dividing process according to the processing procedure stored in the processing procedure storage unit 24, and the result is a divided image data recording unit. 23, and the basic region divided by the basic region feature calculation unit 4 is calculated for each basic region based on the divided image data stored in the divided image data recording unit 23. The calculated basic region features are stored in the basic region feature storage unit 5.

  Here, as the area division process, various area division processes can be used, and a five-dimensional K-average algorithm (Izumi et al .: “Study of a segmentation method using color information and position information together”, electronic information communication The method of extracting brush strokes from the paintings disclosed in the Annual Meeting of the Spring Meeting of the Society of Japan, D680 (1991), and Japanese Laid-Open Patent Publication No. 10-11569 can be used. , Including the position of the basic region on the image and the number of pixels included in the basic region, features that can be extracted from the basic divided region, such as the first principal component and the second principal component, may be used as necessary. A method and state of a typical area division will be described later.

  The basic area feature storage unit 5 includes a RAM, an external storage device, or the like, and stores the basic area features in association with the area number of the basic area in a table format.

  Based on the basic region features stored in the basic region feature storage unit 5, the region integration unit 6 integrates a plurality of basic regions and creates an integrated region according to the processing procedure stored in the processing procedure storage unit 24. The basic area numbers associated with the integrated area number of the integrated area and stored in the integrated area are stored in the integrated area feature storage unit 8 as an integrated area correspondence number table, and the characteristics and The information as to whether or not they have been integrated by the repetitive processing is compiled into an integrated area design map as an integrated area symbol, and is stored in the integrated area design map storage unit 17.

  The integrated region feature calculation unit 7 uses the feature of the basic region integrated into one integrated region in the region integration unit 6 as one unit based on the divided image data stored in the divided image data storage unit 23. calculate.

  The integrated region feature storage unit 8 includes a RAM, an external storage device, or the like, and stores the region features calculated by the integrated region feature calculation unit 7 in a table format in association with the region number of the integrated region.

  The region combining unit 9 includes a rule determined by a processing procedure stored in the processing procedure storage unit 24 from a plurality of adjacent basic regions collected as the same region integration, and an image processing region depending on the case. Using the rules instructed by the instruction unit 18, the basic areas to be combined are selected, and the basic area numbers associated with the combined area numbers of the combined areas are combined as a combined area corresponding number table. Saved in the region feature storage unit 11.

  The combined region feature calculation unit 10 uses the divided image data stored in the divided image data storage unit 23 as a single unit of the basic regions integrated into one combined region by the region combining unit 9. calculate.

  The combined region feature storage unit 11 includes a RAM or an external storage device, and stores the features calculated by the combined region feature calculation unit 10 in a table format in association with the region number of the combined region.

  The partial region dividing unit 12 includes a basic region feature stored in the basic region feature storage unit 5, an integrated region feature stored in the integrated region feature storage unit 8, and a combination stored in the combined region feature storage unit 11. Each region feature is selected from the region features using the rules determined by the processing procedure stored in the processing procedure storage unit 24 and, in some cases, the rules specified by the image processing region instruction unit 18. Each region feature is divided into partial regions based on a method determined in advance by the processing procedure storage unit 24 or a method instructed by the image processing region instruction unit 18, and the partial region and the basic region based on the partial region are integrated. The correspondence relationship between the region and the combined region is stored in the partial region feature storage unit 14 as a partial region correspondence number table.

  The partial region feature calculation unit 13 calculates the feature of the partial region divided by the partial region division unit 12 based on the divided image data stored in the divided image data storage unit 23.

  The partial area feature storage unit 14 includes a RAM, an external storage device, or the like, and stores the features calculated by the partial area feature calculation unit 13 in association with the area numbers of the partial areas in a table format.

  The expression unit 15 includes a basic region feature stored in the basic region feature storage unit 5, an integrated region feature stored in the integrated region feature storage unit 8, and a combined region feature stored in the combined region feature storage unit 11. And the method determined by the processing procedure stored in the processing procedure storage unit 24 and, in some cases, instructed by the image processing region instruction unit 18 from the partial region features stored in the partial region feature storage unit 14 A picture-like image is constructed based on the method.

  The feature changing unit 16 is based on a method determined by the processing procedure stored in the processing procedure storage unit 24 and, in some cases, a method instructed by the image processing region instruction unit 18. 5, the basic region features stored in 5, the integrated region features stored in the integrated region feature storage unit 8, and the combined region features stored in the combined region feature storage unit 11 and the partial region feature storage unit 14. Make selective changes to existing subregion features.

  The integrated region design map storage unit 17 stores an integrated region design map in which the integrated region symbols calculated when the region integrating unit 6 integrates together with the integrated region number, and the integrated region design map is an image processing region. Used for the instruction unit 18.

  The image processing area instruction section 18 is based on the integrated area symbol in the integrated area design map stored in the integrated area design map storage section 17, the area combining section 9, the partial area dividing section 12, the expressing section 15, and the features An operator command input to all or part of the changing unit 16 from a processing instruction preset stored in the image processing area preset storage unit 19 or via a data input unit 20 including a keyboard or the like. To instruct which area is to be processed.

  The image processing area preset storage unit 19 stores a large number of processing instruction presets used by the image processing area instruction unit 18.

  The data input unit 20 is an input device for an operator to send a command to the image processing area instruction unit 18, and includes a keyboard, a mouse or a pen tablet, a voice recognition device, and the like.

  The image data display unit 21 displays a processed image calculated by the expression unit 15 or displays an integrated area symbol stored in the integrated area design map storage unit 17 to an operator by a CRT or a liquid crystal display. It is configured.

  The image data output unit 22 includes a printer that outputs to paper or the like, a display that displays on a CRT or a liquid crystal screen, and displays or prints the processed image calculated by the expression unit 15.

  The divided image data storage unit 23 includes a RAM, an external storage device, or the like, and stores a list of pixels allocated to the basic region by the basic region dividing unit 3.

  The processing procedure storage unit 24 includes a RAM or an external storage device, and stores a processing procedure corresponding to the image input from the image data input unit 1, and the stored processing procedure is selectively processed. It is used in each processing unit (region integration processing, region combination processing, partial division processing, and expression processing).

  In the present embodiment, each unit corresponds to an example of each unit, but the image processing region instruction unit 18 and the image processing region preset storage unit 19 include a region combining unit 9, a partial region dividing unit 12, an expressing unit 15, and features. This corresponds to a part of the changing unit 16.

  Next, an image processing process performed by the image processing apparatus configured as described above will be described. FIG. 2 is a flowchart for explaining the image processing by the image processing apparatus shown in FIG. First, in step S1, image data and a processing procedure are acquired from the image data input unit 1. The acquired image data is sent to the basic area dividing unit 3 and the processing procedure is stored in the processing procedure storage unit 24.

  Next, in step S <b> 2, the basic area is divided in accordance with the processing procedure stored in the processing procedure storage unit 24. Since the division here becomes the basis of the brushstroke that creates the final painting, it is preferable to adaptively change, for example, the initial value of the k-average algorithm in accordance with the characteristics of the original image. For example, when the spatial color change of the original image is severe, there are many features that are regarded as edges, so edges are detected using a known edge extraction operator, and the number of regions is adjusted according to the number of features of the edges. This is a method of changing the initial value of.

An image of region division using the k-average algorithm is shown in FIG. Here, in order to facilitate understanding, a very schematic explanation will be given. First, the captured image data 100A is divided equally into, for example, 3 × 3 areas. Then, area division is performed for one of the areas 100. This is, for example, appropriately divided into nine (in this example, it is the same number as the number of divisions of the image data 100A), and the average colors and centroids of the pixels included in the nine divided areas are calculated and respectively Defined as the center of the region. In the example of FIG. 3A, the respective centroids are shown arranged in one column and three rows for convenience. Then, it is calculated which of the nine area centers is close to all the pixels in the area 100. Here, “close” means that the color and position are close. More specifically, the R (red), G (green), and B (blue) color intensities of each pixel are set to R, G, and B for convenience and the coordinates in the XY coordinates are set to X and Y, respectively. Similarly, for one of the region centers, for example, the region center M11 shown in FIG. 3B, the color intensity is set to r, g, and b, and the coordinates of the center of gravity in the XY coordinates are set to x and y. Perform the calculation.
{(R−r) ² + (G−g) ² + (B−b) ² + (X−x) ² + (Y−y) ² } ^1/2 (1)
That is, with respect to all the pixels included in the region 100, the calculation of the expression (1) is performed with respect to the nine region centers (9 plots in FIG. 3). Then, when attention is paid to a certain pixel, nine calculation results for the pixel are obtained. Among these, the pixel is included in the central region where the minimum value is obtained. If the intensity of the color is the same for both the area center M11 and the other area centers, that is, if the values in the first three terms of equation (1) are the same, the set pixels that are close in distance. It means to become friends with. By repeating such calculation, it is determined to which pixel at the center of the nine regions all pixels are grouped. FIG. 3B shows this state.

  By the way, since the nine area centers are appropriately determined, the following processing is further performed. For each of the grouped regions as shown in FIG. 3B, the center of gravity in the X and Y coordinates is obtained, and the pixel at the center of gravity is set as the new region center. In FIG. 3C, ● represents the center of the old area, and ○ represents the center of the new area. Then, all the pixels in the region 100 are subjected to exactly the same processing as described above (the processing shown in FIGS. 3A to 3C) with respect to the new nine region centers, and the center of gravity is similarly obtained for the obtained group. Find 9 new center areas. When such processing is repeated, the positions of the nine area centers converge. Each of the nine groups at that time is a basic area. The other eight regions other than the region 100 are also grouped in the same manner to obtain a basic region. Note that the image data 100A does not have to be divided vertically and horizontally, but it is better to divide and process the image data in order to reduce the processing time.

When this region division processing is completed, all the pixels belong to the basic region, and therefore, the divided image data describing the corresponding relationship is stored in the divided image data storage unit 23. This divided image data indicates a state in which the image data is divided into basic areas as an image, and a boundary line of the division is described in the image area.
As the data, the number of the basic area is associated with the X and Y coordinate positions of the pixels included in the basic area.

  In the next step S <b> 3, the basic region feature calculation unit 4 calculates the basic region features based on the divided image data stored in the divided image data storage unit 23. When the k-average algorithm is used in step S2, the barycentric value of the basic area, the average color, and the number of pixels included in the basic area have already been calculated at the time when the division process is completed. These also calculate along with the features of the basic region. In addition to these values, the directional element Dn, the flat element Tn, etc. of the area n are calculated by the same method as the technique disclosed in Japanese Patent Application Laid-Open No. 10-11569, and the result is stored in the basic area feature storage unit 5. save.

Here, the direction element Dn represents the direction in which the basic region 300 extends as shown in FIG.
More specifically, the basic region 300 is regarded as an ellipse and indicates the direction in which the long axis extends. For example, the direction parallel to the X axis of the image data is 0 degree, and a predetermined number from 0 degree to 180 degrees Can be equally divided and assigned as the direction element Dn. The direction element Dn can also be determined as the direction of the straight line 201 connecting the pixels by obtaining the pixels at both ends when the variance is maximum. Further, the flat element Tn can be determined as a value obtained by obtaining pixels at both ends when the variance is minimum and dividing the length of the straight line 202 connecting these pixels by the length of the previous straight line.

  FIG. 5 shows a table in which the number of the basic region is associated with the feature of the basic region. For each basic region, the average value of the intensity of each color of red, green, and blue, the position of the center of gravity, A direction element, a flat element, and the number of pixels are described.

In the next step S4, region integration processing is performed based on the values of the basic region features in accordance with the processing procedure instructions stored in the processing region procedure storage unit 24. In the instruction of the processing procedure, constants C ₁ , C ₂ , C ₃ , C ₄ , C ₅ , C ₆ , C ₇ , C ₈ , C ₉ , and C ₁₀ are described. Processing of the integration unit 6 is performed. Details of the processing of the region integration unit 6 will be described with reference to a flowchart. Before that, an overview of the processing will be described with reference to FIGS. 6 to 8 for easy understanding.

What is going to be done in this integration process is to group the basic areas obtained as described above. This grouping is performed by grouping adjacent features that are similar in basic region, but in this embodiment, similar basic regions are once grouped into temporary integrated regions. The temporary integration area is verified and made into a formal integration area. In grouping, whether or not the basic regions are similar is determined by determining whether or not the difference in features of the basic regions is lower than a threshold value. The difference in the characteristics of the basic area is, for example, a difference in color intensity. The intensity of R, G, B in a certain basic area is R, G, B, and the intensity of color in an adjacent basic area is r. , G, and b, the difference in intensity between these colors (characteristic difference D) is expressed by equation (2).
D = {(R−r) ² + (G−g) ² + (B−b) ² } ^1/2 (2)
Alternatively, when taking the addition of the color strength and the above-described direction element as a feature, if the direction element of one basic region and the direction element of the other basic region are Dn and dn, respectively, the difference D of the feature is ( 3) It is expressed by the formula.
D = {(R−r) ² + (G−g) ² + (B−b) ² + (Dn−dn) ² } ^1/2 (3)
Further, as the feature difference, the above-described flat elements may be further added, and it is possible to appropriately determine which one is to be the feature difference. The manner of determining the temporary integration area will be described very schematically with reference to FIGS. Pay attention to each basic region in a predetermined order. If attention is paid to the basic area No. 1, the difference between the features is obtained in order with respect to the basic areas around the basic area No. 1, and if the difference D of the features is smaller than the threshold C6, these basic areas are grouped. Turn into. “Determining feature differences in order with respect to surrounding basic areas” means that the basic area closest to the basic area No. 1 is the top ninth of the basic areas close to each other (this “9th” is applied in an actual device) It is an example that is picked up and compared. In this example, if the basic areas whose feature differences are smaller than the threshold C6 with respect to the basic area 1 are the third and ninth basic areas, the basic areas of the first, third, and ninth as shown in FIG. Are grouped. Similarly, paying attention to the basic area No. 2, they are No. 5, No. 7, and No. 10, which are grouped. If the basic areas 3 and 6 are grouped, the first, third, sixth and ninth are grouped. In the same manner, friends in the basic area are grouped together and integrated. The area integrated in this way is a temporary integration area, and is shown in FIG.

  This temporary integration area may be used as the integration area as it is, but this temporary integration area is verified in order to obtain an advantage that more detailed processing can be performed later. Assuming that processing (grouping check) is performed for all the basic areas as a temporary integration area, not all the basic areas belong to one of the temporary integration areas. We will consider it as a temporary integration area.

  After the first grouping check is thus performed, the same processing is performed again for all the basic areas. That is, the second grouping check is performed. However, the value of the threshold C6 at this time is set higher than the first threshold. Grouping when the feature difference is smaller than the raised threshold means that the grouping condition is relaxed. When grouped in this way, the first grouped items do not leave each other, and a new basic area is added to the group. FIG. 8B shows the progress of the second temporary integration process. Here, since the basic area 5 is included in the temporary integration areas 1 and 2, the temporary integration areas 1 and 2 are integrated into the smaller number 1 and 2 is lost. FIG. 8C shows the second temporary integration processing result. Then, it is checked how much the number of basic areas included in each temporary integration area has increased from the first time. For example, in FIG. 8, the number of basic areas with the temporary integrated area number 1 is 4 for the first time and 10 for the second time, and the increase rate is 15%. If the rate of increase in the number of basic areas is less than or equal to a predetermined value, for example, if it is 10% or less, the temporary integrated area is formally designated as an integrated area. This is because it is judged that it is a group with a strong degree of characteristics compared to the surroundings, because it gathers friends under a certain condition, and the number of friends does not change so much even if you loose the condition a little and gather friends. is there. In more detail, as shown in a later flowchart, it is added to the condition that the number of basic areas in the temporary integration area is not less than a predetermined value. The basic area group that has become a formal integrated area is excluded from the subsequent grouping process.

  If such a grouping process is performed a predetermined number of times, for example, 7 times, it is recorded at which stage it has been upgraded to a formal integrated area. For example, if the temporary integration area number j becomes a formal integration area in the second grouping process, it is a second-stage temporary integration area, and in the following description, this is called a 2th uniform integration area. In some cases. Eventually, a 2th uniform integrated region, a 3th uniform integrated region,..., A 7th uniform integrated region are formed. It can be said that the group that has been integrated quickly and becomes a uniform integrated region is a group with a strong degree of feature to the surroundings. As will be described later, when the integrated areas are prioritized and managed in this way, fine brushstrokes can be expressed.

  In this example, in addition to the uniform integrated region, a contrast integrated region is also handled. After completing the i-th grouping process, for example, the second grouping process, it is checked whether or not the temporary integrated area is a formal uniform integrated area. If it is determined as “No”, the temporary integrated area It is determined whether or not it is a region having a strong contrast, and if it is determined that the contrast is strong, it is managed as a contrast integrated region and deviates from the subsequent grouping process. Also in this case, the stage at which the comparison integration process is performed is recorded, and the 2th comparison integration area, the 3th comparison integration area,..., The 7th comparison integration area is formed.

  Based on the schematic description above, specific steps of the integration process will be described with reference to the flowcharts shown in FIGS.

  First, the processing integration count i of the region integration unit 6 is set to i = 0, and the processing count i is increased in step S101 shown in FIG.

Next, in step S102, it is determined whether or not the number of repetition end times C ₁ determined in advance, for example, C ₁ = 10 is greater, and if so, the process ends at that point. If it is smaller, the process proceeds to S103.

  In step S103, a temporary integration area is determined. The process of determining the temporary integration area will be described later with reference to the flowchart of FIG. When the i-th temporary integration area is determined, it is checked whether all temporary integration areas satisfy the conditions of the next step in the order of serial number j. J = 1 is set as an initial value (S104).

  In step S105, it is checked whether the following conditions are satisfied.

N _{i, j} ≧ G / (U × i + 1) and (N _{i, j} −N _{i−1, j} ) / N _{i−1, j} <C ₂
Here, N _{i, j} is the number of basic fee areas included in the j-th temporary integration area obtained for the i-th time, G uses the total number of basic areas, and U is described in the processing procedure. For example, 120 is used. Based on this condition, it is determined whether or not basic areas larger than a predetermined number are gathered in the temporary integration area.

Further, another formula is used to check whether or not the number of basic areas included in the temporary integration area corresponding to the (i-1) th time is largely changed in the i-th temporary integration process. Here, C ₂ also is a value described in the procedure, for example, a 1.2. As described in the schematic description above, the temporary integration region is set to the i-th time by making the threshold C _6i used in the i-th temporary integration processing larger than the threshold C _6i-1 used in the i- _1th integration. The number of basic areas included in the temporary integration area created by the temporary integration process is increased. In this case, it is examined that this increase is 1.2 times or less.

  When the above two conditions are satisfied, the process proceeds to step S107, where the temporary integrated area is officially determined as the uniform integrated area, and the basic area is excluded from the integration targets. That is, when the number of times of grouping basic areas is increased by 1, the rate of increase in the number of basic areas included in a group is greater than or equal to the set value, and the number of basic areas in the group is greater than or equal to the set value. When that happens, make the group a formal integration area.

  Next, the process proceeds to step S108, where the determined uniform integrated region is named i-th uniform integrated region, which is stored as an integrated region symbol together with the integrated region number through the integrated region design map, and further determined uniform integrated region. And the correspondence table of the numbers of the basic areas belonging thereto are stored in the integrated area feature storage unit 8.

Returning to step S105, if the above two conditions are not satisfied, the process proceeds to step S106. In step S106, by checking whether all three conditions of N _{i, j} ≧ C ₃ , N _{i, j} ≦ C ₄ , and A _{i, j} > C ₅ are satisfied, the contrast strength of the integrated region is increased. Deciding. Here, A _{i, j} is a value determined as follows. As shown in FIG. 11, for example, a basic region that is included in the focused integrated region j and touches the boundary 300 is k, and a basic region that touches the boundary 300 of the integrated region j and is not included in the focused integrated region j. The basic area closest to the basic area k is k ｀. Then, if the difference between the brightness of the basic region k and the lightness of the basic region is k こ の, and this difference is A _{i, j, k} , all included in the focused integrated region j and in contact with the boundary 300 A _{i, j is} an average value of A _{i, j , k} obtained by performing the same processing on the basic region k.

The C _5, for example, the brightness range of the original image is L, if the reference threshold was B = L / 80, C ₅ can be used six times the value of the reference threshold B. Note that the reference threshold is a unit value computer has to determine values such as C _5. However, C ₅ is not limited to six times the reference threshold B, it is possible to use any value within the 20-fold from 1 times. Further, by using C ₃ = 2 and C ₄ = 5, an integrated region having a strong contrast and a small area is selected.
Even if the brightness contrast with the surroundings is the same, when i is a small value, there is little variation in the brightness of the basic area in the integrated area, so that a stronger contrast can be felt. In this step, _{Ai, j} only deals with brightness contrast, but instead of brightness contrast or in addition to brightness contrast, formulas that consider chromaticity and saturation should be adopted. You can also. If all three conditions are satisfied, the process proceeds to step S109. If not satisfied, the process proceeds to step S111. The value of j is incremented by one, and the process proceeds to step 112.

  When the process moves to step S109, the temporary integration area is formally determined as the comparison integration area and excluded from the integration target. In the next step S110, the i-th contrast integrated region is recorded in the integrated region design map, and then the process proceeds to step S111, the value of j is incremented by one, and the process proceeds to step S112.

  In step S112, the value of j is compared with the total number of temporary integration areas of the i-th process. If the values are the same, the transition process to the formal integration area has been completed for the temporary integration area at this stage. If there is a difference, the process of transition to the formal integration area remains, so the process proceeds to step S104.

  Next, the method for determining the temporary integration area in step S103 will be described using the flowchart shown in FIG. This flow corresponds to FIGS. 8A to 8C. Here, it is assumed that the determination flow is for the i-th temporary integration region. In step S201, for the temporary integration area number j, first, the first basic area j 'included in the temporary integration area number j is determined, and the basic area j' is entered into the list of temporary integration area numbers j.

Subsequently, in step S202, a feature difference Dj ″ between the noticed basic region j ′ and the surrounding basic region j ″ is calculated. In step S203, Dj ″ <C _6i is checked. If this condition is satisfied, the process proceeds to step S205, and the basic area j ″ is added to the list of temporary integration areas j. _Here, the value of _{C 6i} is a _{C 6i = i × B / C1} , B is a reference threshold B shown above. For this reason, since the value of C _6i increases as the number of times i increases, basic regions having large color differences are also easily integrated each time the number of integrations increases.

If the condition of Dj ″ <C _6i is not satisfied, step S204 is skipped and the process proceeds to step S205. In step S205, it is checked whether or not there is an uncalculated basic region j ″ around the noted basic region j ′. If there is, the process proceeds to step S202, and if not, the process proceeds to process S206. Transfer.

  In step S206, it is checked whether or not there is a basic area j 'that is not focused in the temporary integrated area list j. If there is, the process proceeds to step S202, and if not, the process proceeds to step S207. In step S207, it is checked whether all the basic areas belong to the temporary integrated area. If they belong, the temporary integrated area processing is terminated, and the process returns to step S103 in FIG. If there is something that does not yet belong, the temporary integrated area number j is incremented by one, and the process returns to step S201. This is the flow of the process for determining the temporary integration area shown in FIG. 4, and the integration area is determined by repeating the processes of FIGS. 3 and 4.

  As a feature of this integration process, if the number of basic areas included in the temporary integration area does not change even if the temporary integration is changed again by changing the threshold, or if it does not change much, it is officially determined as the integration area, for example, In the landscape image, there is an effect that it is possible to prevent the integrated region from being interrupted at a place in the middle of gradation such as the sky. In order to grasp the degree of change in the number of basic areas included in the temporary integration area, it is only necessary to have a preset value for the rate of increase in the number. If the setting value is 1, provisional integration is performed again. Corresponds to the case where the value does not change, and if the set value is close to 1, it means that the value has not changed much. On the other hand, if the first temporary integration area is treated as a uniform integration area, priority cannot be given to the integration areas, so the degree of freedom of expression is narrowed. The arrangement may be unnatural, and the present invention includes such a case.

  The integration process ends in step S4 in FIG. 2. At this point, the correspondence between the integrated area number and the basic area (which basic area is included in the integrated area of which integrated area number). Is stored in the integrated region feature storage unit 8, and each integrated region is integrated by the i-th integration process and is written in the integrated region design map as a uniform integrated region or a contrast integrated region. Become.

Further, when performing this integration process, if the flat element Tn is one of the basic region feature n meets Tn> C _7, removing the base area from the integration processing, edge integration region the base region “Edge” may be written in the integrated area symbol of the integrated area design map.
Further, a basic area that is not included in any integrated area after the repetition process is completed may be named an independent integrated area and written as an integrated area symbol in the integrated area design map. In this way, by adding an edge integration area consisting of one unintegrated basic area and an independent integration area, all the basic areas are written in the integrated area design map as an integrated area symbol of the integrated area. It will be. Here, for example _{C 7} may be used 0.95. The edge integration region and the independent integration region often play important expressions in the image, and correspond to, for example, the contour line between the human face and the background. Blur can be avoided.

  Returning to FIG. 2, in step S5 in FIG. 2, the pixels grouped as the integrated region are collected from the divided image data in the divided image data storage unit 23, and the integrated region feature is calculated. The items to be calculated are the same items as in step S3, that is, the barycentric value of the integrated region, the average color, the direction element Dn of the integrated region n, the flat element Tn, and the number of pixels included in each basic region belonging to the integrated region. These features of the integrated region are stored in the integrated region feature storage unit 8.

  Next, in step S6, when processing is automatically performed using the information of the processing preset stored in the image processing region storage unit 19, the integrated region design map is displayed on the image data display unit 21, and the map is referred to. The user can select one of the two when the user instructs the processing using the data input unit 20.

  In FIG. 2, the processing from step S9 to S13 is described sequentially, but if there is no description of these processing in the processing procedure stored in the processing procedure storage unit 24, these steps are skipped. In some cases, the process proceeds to step S13 in which the image expression unit 15 calculates image features. In this case, a painting work is created from the basic area feature and the integrated area feature. When creating a touch from these two area features, a small brush is created from the basic area feature and a large brush is created from the integrated area feature. . Until now, it was possible to make a wide range of one color by painting the same color on many small brushes, but such extremely different brush sizes are not created simultaneously in one picture. It was.

  The processing from step S9 to step S13 is for adding various expressions to a painting to be created by combining and enlarging the basic regions, or conversely dividing and reducing the processing. It is also possible to describe in the processing procedure stored in the storage unit 24 that only one of steps S9 and S10 related to the combination processing and steps S11 and S12 related to the partial region processing is performed.

  That is, whether the integrated area obtained by the processing up to step S5 is expressed (drawn) using it as it is, or some basic areas in each integrated area are combined (steps S9 and S10). Then, it is selected whether drawing is performed in units of these combined areas, or whether each integrated area is partially divided and drawn in units of divided areas (steps S11 and S12). Note that the final representation processing (drawing processing) includes the case where the basic region is used as a unit in addition to these region units.

  The following describes how step S9 is executed. When automatic processing is selected by the image area instruction unit 18 in step S6, the image area instruction unit 18 performs preset processing from the image processing area preset storage unit 19. Read instructions. The preset processing instruction is, for example, for the uniform integrated regions from th1 to th7, those having high priority, for example, the uniform integrated regions of th1 and th2, are subjected to region combination processing, and the uniform integrated region of th3 is divided into partial regions. This is an instruction to perform the process, and neither process is performed for the uniform integrated region from th4 to th7. The example shown here is merely a schematic explanation, and these instructions are set in advance as appropriate.

  When manual processing is selected in step S6, the integrated design map stored in the integrated design map storage unit 17 is displayed on the screen as shown in FIG. 12, and the operator operates as described above while viewing this screen. Manually input such instructions. Therefore, the manual preset processing can usually express more finely, but automatic processing may be selected if importance is attached to simplification of operation.

If step S9 needs to be executed for a certain integrated area, first, C ₈ basic areas in the integrated area are paired. As a rule for pairing, for example, in the case of C ₈ = 2 it is preferable that two adjacent ones are connected as vertically as possible, but a rule that makes the connected result close to a circle can also be used. . The rule that two adjacent ones are connected vertically is a rule that searches for a basic region at a center of gravity close to the center of gravity of the basic region of interest, and connects those whose direction vectors of the basic regions are close to a straight line. It is done.

  In step S10, the combined region feature is calculated using the region combined in the previous step S9 from the pixel of the divided image data in the divided image data storage unit 4 as one region, and the calculated combined region feature is used as the combined region feature storage unit. 11 to store. The combined region feature here is the same as the feature described above shown in FIG.

Next, when step S11 needs to be executed for a certain integrated area, one of the basic area, the integrated area, and the combined area that has already been calculated is selected according to an instruction from the image processing area instruction unit 18. The region is divided into C ₉ partial regions, the C ₉ region features are calculated from the pixels of the divided image data in the divided image data storage unit 4, and the calculated combined region feature is combined with the combined region feature storage unit. 14 stored. FIG. 13 shows an image of the division of the partial area. For example, only P3 of the integrated areas P1, P2, and P3 is vertically divided into, for example, three. As the number of divisions, “3” is a numerical value for convenience. Thus, the rule for dividing into partial areas is also determined in advance, and the area to be partially divided is divided into partial areas based on the rule. This division processing is to make it appear as if it was drawn with a brush when expressed.

  Here, the process will be described assuming that the above steps S9 to S12 are used. Step S6 will be described first when it is not automatic processing. FIG. 12 shows an example of the integrated area symbol in the integrated area design map displayed on the image data display unit 21. The area filled in black is the edge integration area, the white part is the contrast integration area and the independent integration area, and the part of the texture shown as uniform is uniformly integrated when the second temporary integration area is obtained. Area determined as an area, 3 uniform, 4 uniform, 5 uniform, and 6 uniform are the uniform integrated areas obtained in order, and 7 uniform is a summary of the areas determined as the uniform integrated area at the 7th time or more. is there. Since the sky with gradation is grouped into one, and the part is labeled as 2 uniform, it can be seen that the integration process of the proposed method worked well and the sky was integrated into one.

  The edge integration region, the contrast integration region, and the independent integration region are drawn with the brush created from the basic region, and the uniform integration region determined with a small number of repetitions is the integration region or the processing of the previous step in steps S9 and S10. It means that it is preferable to create a touch based on the coupling area to be created and the partial divided areas created in step S11 and step S12. The problem here is that there is a variation depending on the type of image in which group these uniform regions integrated with a large number of repetitions should be put.

  The uniform integrated area can be automatically processed if it is decided to make all the areas for adding expression. However, when aiming for a higher quality work, the processing is performed manually in the image data display unit 21. After looking at the integrated area symbol in the integrated area design map shown, a processing instruction can be given in units of the integrated area or basic area.

  As a result, in step S7 or step S8, an instruction as to which stroke feature is to be made for each integrated area is issued, so that subsequent processing steps S9, S10, S11, and S12 are executed according to the instruction. Is done. That is, the integrated area is a set of basic areas, and there may be a combined area and a partial divided area in the integrated area. Assuming that the processing instruction in step S7 or S8 is the instruction that “the unit area of expression is an integrated area” for the 3th uniform integrated area No. 10, the integrated area 10 is expressed as a unit (the integrated area 10). Make brushstrokes from features). In this case, for the integrated area 10, colors (for example, R, G, B) and their brightness are designated for all the pixels included in the integrated area 10 based on the characteristics of the integrated area 10. It will be. Also, if the processing instruction in step S7 or S8 is an instruction “use unit area of expression as basic area”, each basic area included in integrated area 10 is expressed as a unit (from basic area characteristics to brush strokes). Will be made). In this case, for each basic area, a color is designated based on the characteristics of each basic area for all pixels included in the basic areas. Furthermore, if the processing instruction is an instruction that “the unit area of expression is a combined area”, the combined area included in the integrated area 10 is drawn as a unit of expression, or the processing instruction is “expression. If it is an instruction that “the unit area is a partial divided area”, the partial divided area included in the integrated area No. 10 is drawn as a unit of expression.

  In addition, since each integrated area is stored in the integrated area design map storage unit 17 as to how many times (in what stage) the temporary integrated area has become a formal integrated area, it is not a feature of the integrated area. The basic area included in the integrated area is used as a unit area for expression, which stage of integrated area is used, and the characteristic area of the integrated area is used as a unit area for expression. Data is stored in memory. For example, for a uniform integrated region of 1th to 3th, for example, a feature of the integrated region integrated region is a unit region of expression, and for a uniform integrated region of 4th to 7th, a basic region included in each integrated region is expressed. An example in which information such as a unit area is previously stored in a memory can be given. The numerical value at this stage can be selected via the input unit, and in this way, it is possible to automatically perform an operation of which region feature is to be used for each integrated region. That is, in the present invention, such work can be performed manually or automatically.

  In step S13, the characteristics of each area feature that has already been calculated, more specifically, the characteristics of the area (basic area, integrated area, combined area, or partially divided area) that has become the above-described expression unit can be changed. This processing step is also executed only when there is a description in the processing procedure of the processing procedure storage unit 24. Part of the processing performed here (changing the shape and direction of the area) is the same as the technology constituting the group feature processing unit disclosed in Japanese Patent Laid-Open No. 10-11569, but the basic part is the same. The feature processing disclosed in the 11669 publication applies only to region features described in this patent as basic regions. On the other hand, in this patent, various region features (basic region, integrated region, combined region, or partial division region) are already described in the same format as shown in FIG. The region feature can be changed in a state including the.

  Further, a range in units of basic areas to which this process is applied is also selectively instructed from the image processing area instruction unit 18. In addition, each area feature that has been changed based on the difference in color between the integrated area in which attention is paid to colors in this step and the integrated area around it is returned to each area storage unit.

  FIG. 14 shows a detailed block diagram of the expression unit 15, and FIG. 15 shows details of the processing in step S14. Step S14 will be described with reference to FIG.

  In step S301, for each integrated region, the integrated region symbols described in the integrated region design map are read, and according to the instruction from the image processing region instruction unit 18, the types of regions corresponding to the integrated region symbols are collected and assigned serial numbers. Modify the expression area set. In other words, when it is finally expressed, it is only necessary to specify which area is the unit of expression, and a series of processes such as basic area integration, basic area combination in the integrated area, and partial division of the integrated area are essential. This is done in order to increase the degree of freedom of selection of the area that becomes the expression unit, and it is only necessary to specify the area that becomes the unit when expressing. For example, if the integrated area is partially divided, it is only necessary to number each divided area and to understand the area characteristics of each area. It does not matter whether the basic area belongs.

  In step S14, when an expression unit is determined, a numbering is newly added to the unit area, and each numbered area and its features are extracted into a table. This table corresponds to the above expression area set. For example, in the modeled explanation, for example, for one image, as a result of processing, three partial divisions in the basic area 2, 5, 10, the integrated area 2, 3, and the integrated area 4 Assuming that it is composed of areas, each of these areas is newly numbered, and at this stage it is no longer relevant what kind of area it is.

  In addition, adding a pictorial expression to a highly contrasted part or an edge part that is an outline destroys important features of the original image. For example, if you are a human face, you do not know who you are drawing. It will disappear. In order to avoid this, the image processing area instruction unit 18 includes an integrated area symbol such as an edge integrated area, a contrast integrated area, an independent integrated area, and a uniform integrated area determined by a large number of repetitions. Should be drawn in a realistic expression that does not add much expression, so it is preferable to instruct to add the basic area included in the integrated area as an expression unit to the expression area set rather than using the integrated area as an expression unit. . On the other hand, it is preferable to instruct the integrated region itself, the combined region, and the partial region to be added to the expression region set as an expression unit so that a painting-like expression can be added to the uniform integrated region determined with a small number of repetitions.

  Steps S302 to S304 to be performed next are processing for determining which expression unit region each pixel included in the image belongs to. This processing calculates the distance to each representation unit area in order for each pixel included in the image, for example, determines the representation unit area physically closest to the center of gravity position as the representation unit area to which the pixel belongs. However, doing so increases the load on the computer for distance calculation, so a geograph table is used here. This geograph table is a table that lists, for each expression unit area, expression unit areas that are physically close to the center of gravity of the expression unit area in order from the nearest number. By looking at this, it is possible to find an expression unit region existing in the vicinity of an arbitrary expression unit region.

  For the first (k = 1) pixel in the pixel group included in the image, a predetermined distance calculation is sequentially performed for a plurality of expression unit regions located in the vicinity of the pixel, and the expression having the minimum distance is obtained. The unit area is written in the pixel feature storage unit 28 together with the pixel number. A plurality of distance calculation formulas are prepared in the distance calculation formula storage unit 26, and are calculated using a distance calculation formula selected from them. In a certain calculation formula, even if the position of the pixel k and the center of gravity of the representation unit region n are closest, the pixel is on an extension line in the extending direction of the representation unit region n ｀ that is farther from the representation unit region n. If k is located, a value closest to the expression unit region n 表現 is obtained for the pixel k. That is, in this case, the expression unit area is emphasized in the direction of extension, and a distance calculation formula may be selected according to the expression method.

  For the second and subsequent pixels k, the distance calculation may be performed in order for the expression unit area to which the pixel (k-1) belongs and the neighboring expression unit area indicated by the geograph table, and the expression with the smallest calculated value. Judge as belonging to the unit area. This has the advantage that the load for distance calculation is reduced.

  Returning to the description of the flow of FIG. 14 based on such an outline, in step S302, a geograph table is created using the expression region set. As a method for creating a geograph table, the method disclosed in Japanese Patent Application Laid-Open No. 10-11569 can be used. However, in short, it is only necessary to table expression unit regions that are close to each other in position. Therefore, an algorithm for creating a table may be created as appropriate.

  In step S303, the pixel feature calculation unit 25 reads, for example, other representation unit regions listed in the representation unit region to which the k-1th pixel belongs, from the geograph table, by representing the pixel number as k, for example. The distance to the target pixel k is calculated using the distance calculation formula selected from the distance calculation formula storage section 26 in accordance with the processing procedure of the storage section 24 and the instruction of the image processing area instruction section 18. In the case of k = 1, the calculation is performed for the neighborhood region of the expression unit region 1. As the distance calculation formula, formula (8) disclosed in Japanese Patent Laid-Open No. 10-11569 can be used, but not limited to this, various calculation formulas are prepared in the distance calculation formula storage unit 26 in advance and integrated. The distance is calculated by selecting according to the integrated area symbol described in the area design map, and the number of the expression unit area having the smallest distance calculation value is written in the pixel feature storage unit 27 together with the pixel number in step S304. As a result, various brushstrokes can be created in one processed image.

  When the expression unit area is composed of partial areas, it is often seen by impressionists because a small brush corresponding to the partial area can be created by calculating the distance for each partial area. A touch of a blurred brush can be expressed. In addition, by using the values of a plurality of partial areas as a group, it is possible to create one more complex brush shape.

  For example, when three partial areas are made from the center and both ends of one basic area, a quadratic curve passing through the center of gravity of the three partial areas can be defined. In the disclosed distance calculation formula (8), the first principal component of the region corresponds to the local Y axis of the region. The distance calculation is performed after the Y axis is coordinate-converted into the above-described quadric curve. Then, the brush will bend along a quadratic curve.

  By using the technique of bending the brush stroke using the center of gravity of this partial area, if the original area is bent, it can be approximated by the brush stroke of a quadratic curve. If the entire processed image is accurate and more expressive, it is preferable to obtain partial regions of all regions in step S12.

  In step S305, the pixel feature conversion unit 28 selects one of a plurality of calculation formulas stored in the pixel feature conversion formula storage unit 29 in accordance with the processing procedure of the processing procedure storage unit 24 and the instruction of the image processing area instruction unit 18. Then, the color of the pixel is calculated from the feature of the expression unit area of the number stored in the pixel feature storage unit 27 of that pixel, and the value is written in the corresponding pixel position of the processed image. By doing so, it becomes possible to give a color change in one brush stroke. Since each pixel of the image captured by the computer in step S1 in FIG. 2, for example, each pixel of the photographic image belongs to which expression unit region belongs in the processing up to step S304, in step S305, the pixel is A color is assigned to the pixel based on the feature corresponding to the expression unit region to which it belongs. According to the simplest method, for example, the average colors of R, G, and B in the expression unit area are assigned. In this example, a predetermined calculation process is performed on the feature, and the average color is assigned based on the calculation result. I have to. The calculation process is a process in which the average color decreases as the pixel moves away from the barycentric position of the expression unit region.

Step S306 to increase one pixel numbers in the value of k is checked is greater than _{C 10} in step S307, if _{k> C 10,} the process ends, to meet the condition and returns to step S303 Then, the processing for the next pixel is continued.

  When the process in FIG. 15 is completed, the process returns to step S15 shown in FIG. 2. In step S15, the processed image calculated in step S14 is output from the image data output unit 22 and printed or saved. The image data display unit 21 can also display it for confirmation. That is, since the color is assigned to each pixel up to step S14, in this step S15, each pixel is colored based on the assigned color, and thereby, for example, a photographic image captured earlier is processed to produce a painting style. A processed image in which the brush stroke is expressed is obtained.

  FIG. 16 shows the divided image data stored in the divided image data storage unit 23. Each usage area shown here is a basic area.

  FIG. 17 shows the basic region feature, the combined region feature, and the divided region feature calculated from the divided image data, the edge region, the isolated region, the contrast region, and the region determined as the unified region by repeating the process six times or more. The region of the feature and the pixel of interest, and the region determined as the unified region by repeating 2-5 times, is a processed image created by calculating the distance between the partial region feature and the pixel of interest.

  FIG. 18 illustrates basic region features, combined region features, and divided region features calculated from divided image data, and edge regions, isolated regions, and contrast regions are determined as basic region features and target pixels, and a unified region. The coordinate conversion is performed on the region with the quadratic curve calculated using the centroid value of the partial region feature calculated from the combined region, and further, the region is moved along the coordinate-converted first principal component axis. Only the line width proportional to the number of included pixels is a processed image created by inserting a conspicuous curve.

It is a block diagram which shows the structure of the image processing apparatus by one Embodiment of this invention. 3 is a flowchart for explaining image processing by the image processing apparatus shown in FIG. It is explanatory drawing which shows a mode that image data is divided | segmented into a basic region. It is explanatory drawing for demonstrating the direction element which is one of the characteristics of a basic region. It is explanatory drawing which shows an example of the characteristic of a basic region. It is explanatory drawing for demonstrating a mode that a basic area | region is integrated and a temporary integration area | region is formed. It is explanatory drawing for demonstrating a mode that a basic area | region is integrated and a temporary integration area | region is formed. It is explanatory drawing which shows a mode that the basic area | region contained in a temporary integration area | region increases by repeating trial. 3 is a flowchart for explaining an integrated region creation process shown in FIG. FIG. 4 is a flowchart for explaining a temporary integration area creation process shown in FIG. 3. FIG. It is explanatory drawing which shows a mode that a contrast integrated area | region is formed. It is a figure which shows an example which displays the integrated area symbol described in the integrated area design map as an image. It is explanatory drawing which shows a mode that an integrated area | region is partially divided. It is a block diagram for demonstrating in detail the structure of the expression part shown in FIG. 3 is a flowchart for explaining calculation of pixel features shown in FIG. 2. It is a figure which shows an example of division | segmentation image data. It is a figure which shows an example of a process image. It is a figure which shows an example of a process image.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image data input part 2 Image feature extraction part 3 Basic area division | segmentation part 4 Basic area feature calculation part 5 Basic area feature storage part 6 Area integration part 7 Integrated area feature calculation part 8 Integrated area feature storage part 9 Area | region coupling | bond part 10 Connection area | region Feature calculation unit 11 Combined region feature storage unit 12 Partial region division unit 13 Partial region feature calculation unit 14 Partial region feature storage unit 15 Representation unit 16 Feature change unit 17 Integrated region design map storage unit 18 Image processing region instruction unit 19 Image processing region Preset storage unit 20 Data input unit 21 Image data display unit 22 Image data output unit 23 Division image data storage unit 24 Processing procedure storage unit 25 Pixel feature calculation unit 26 Distance calculation formula storage unit 27 Pixel feature storage unit 28 Pixel feature conversion unit 29 Pixel feature conversion storage unit

Claims (25)

A basic area dividing processing means for grouping each pixel with at least a color as one of the determination criteria for the pixels included in the image data, thereby dividing the image data into a plurality of basic areas;
Basic area feature calculating means for calculating the characteristics of the basic area including at least color features for each divided basic area;
Based on at least the color of the basic region features calculated by the basic region feature calculating means, region integrating means for integrating a plurality of basic regions to produce an integrated region;
For each integrated region, an integrated region feature calculating means for calculating a feature of the integrated region including at least a color feature based on pixel data included in the basic region;
An image processing area instruction unit that determines whether the integrated area is a unit area of expression or a basic area included in the integrated area is a unit area of expression according to the integrated area;
An image comprising: an expression unit that determines a color of a pixel included in a unit area based on a feature of the unit area of the expression for each unit area of the expression determined based on the selected result Processing equipment. When the most preferable region feature is selected by using the two region features of the integrated region feature and the basic region feature, or one of the region features and pixel position information as an input value, the expression unit has a minimum value. In addition to the basic region feature of the basic region and / or the integrated region feature of the integrated region that gave the smallest value as a result, the minimum value of the previous calculation result is calculated. 2. The image processing apparatus according to claim 1, wherein pixel features such as color information are calculated by using all or some of values calculated from a difference between the value and the second smallest value as input values. An area combination means for selecting and combining two or more basic areas from a plurality of basic areas integrated in the same integrated area;
A combined area calculating means for calculating the characteristics of the combined area combined by the area combining means,
The image processing area instruction unit determines whether the integrated area is a unit area for expression or a basic area included in the integrated area is a unit area for expression, depending on the integrated area. The image processing apparatus according to claim 1, wherein the combined region is determined as a unit region of expression. Partial region dividing means for dividing the integrated region and / or the basic region into a plurality of partial regions,
A partial region feature calculating means for calculating a partial region feature from each region feature with reference to the integrated region feature or the basic region feature;
The image processing area instruction unit determines whether the integrated area is a unit area for expression or a basic area included in the integrated area is a unit area for expression, or a part obtained by dividing the integrated area according to the integrated area The image processing apparatus according to claim 1, wherein the image processing apparatus determines whether the area is a unit area of expression. The representation means includes an integrated area feature of the integrated area stored in the integrated area feature storage means, a basic area feature stored in the basic area feature storage means, and a combination stored in the combined area feature storage means. In addition to these three types of region features, or in addition to these, four types of region features from among the partial region features calculated by the partial region dividing means, or any of several types of region features therein The basic region feature of the basic region that gave the minimum value as a result of performing a predetermined calculation 2 that is considered to indicate the minimum value when the most preferable feature is selected with the pixel position information as an input value, Or, it is calculated from the difference between the integrated region feature of the integrated region, the combined region feature of the combined region, or the minimum value of the calculation result of the predetermined calculation 2 and the second smallest value. Image processing apparatus as claimed in any one of the 4, characterized in that calculating the pixel feature of the pixels constituting the image as input some of all or therein, of the values. The image feature extraction unit includes three types of region features, ie, an integrated region feature, a basic region feature, and a combined region feature, or in addition to the four types of partial region features calculated by the partial region dividing unit. Type region features, or any number of each type of region feature,
A feature changing means for changing each area feature based on a value calculated from each area feature of the basic area, the integrated area, the combined area, or the partial area located around the basic area, the integrated area, the combined area, or the partial area. And an image processing apparatus according to any one of claims 1 to 5. The area integration means finds a basic area without a mark and marks it with a mark 1. Mark a basic area where the difference in feature quantity between the basic areas adjacent to the basic area is smaller than the first threshold value, and further change the feature quantity between the basic areas adjacent to the marked basic area. Repeat the process of marking 1 even for basic areas where the difference is less than the first threshold. Finally, the basic areas marked with mark 1 are combined into a temporary integrated area 1. Find a basic area that is not marked, add a mark 2 to the basic area, repeat the same process, and finally repeat until there are no unmarked basic areas. Next, paying attention to a temporary integration region in which the number of basic regions included in the temporary integration region is larger than the first number, the difference in feature amount between the adjacent basic regions is larger than the first threshold value, and the second threshold value. A smaller basic area is grouped in the same way, and the new temporary integrated area that has been collected includes the basic area that was included in the temporary integrated area that was noted earlier. When the value obtained by dividing the number of areas by the number of basic areas collected in the temporary integrated area of interest is larger than 1 and smaller than the specified value 1, the temporary integrated area is formally determined as an integrated area. The image processing apparatus according to claim 1, wherein the image processing apparatus is characterized in that: After the second temporary integration area is determined, the area integration means sets the third temporary integration area and the second temporary integration area formally using the second threshold value as the first threshold. After that, when the same operation is repeated several times, the integration process is continued only for the remaining basic areas excluding the basic area determined as the integrated area in the previous process. When the determination process as the nth formal integration area is completed, a serial number is assigned to the integration area, and the serial number and the nth uniform integration area are recorded in the integration area design map. The image processing apparatus according to any one of claims 1 to 7. The area integration means pays attention to a temporary integration area whose basic area number is larger than the first number and smaller than the second number after the n-th repetitive processing, and for the basic area located near the boundary of the temporary integration area. Among them, the absolute value of the average value of the difference between the basic area feature of the basic area in the temporary integration area of interest and the basic area feature of the basic area adjacent to and outside the temporary integration area of interest is the prescribed value 2 If it is larger, a serial number is assigned to the temporary integration area, the serial number and the nth comparison integration area are recorded in the integration area design map, and after this temporary integration area is excluded from the integration processing target, n + 1 The image processing apparatus according to claim 1, wherein a repetitive process is performed for the first time. The region integration means excludes a basic region having a basic region feature larger than the specified value 3 in advance before performing all processing, and describes the information of the excluded basic region in the previous integrated region design map. The image processing apparatus according to claim 1, wherein the image processing apparatus is provided. The area combining means includes a basic area in which the number of basic areas to be combined as a single combined area can be combined with a plurality of adjacent basic areas collected in the same integrated area by the area integrating means. The image processing apparatus according to claim 1, wherein a relationship with a feature is changed according to information described in an integrated region design map of an integrated region to which the basic region belongs. The representation means includes an integrated area feature of the integrated area stored in the integrated area feature storage means, a basic area feature stored in the basic area feature storage means, and a combination stored in the combined area feature storage means. In addition to these three types of region features, or in addition to these, four types of region features from among the partial region features calculated by the partial region dividing means, or any of several types of region features therein Are selected based on the information described in the integrated region design map of the integrated region in which each is included, and further, pixel features are selected from the features selected based on the information described in the integrated region design map. The image processing device according to claim 1, wherein the pixel feature is calculated by switching a calculation formula to be calculated. The partial area dividing means divides an integrated area, a basic area, or a combined area into a plurality of partial areas, and for each partial area, an integrated area feature of an integrated area stored in the integrated area feature storage means. Or when the partial region feature is calculated by the partial region feature calculation unit with reference to the basic region feature stored in the basic region feature storage unit or the combined region feature stored in the combined region feature storage unit Depending on the information described in the integrated region design map of the integrated region to which the partial region belonged, a different calculation formula is selected, and the partial region feature of the partial region is calculated using the calculation formula. The image processing apparatus according to claim 1, wherein a feature is stored in the partial region feature storage unit. The feature change means is stored in the integrated area feature of the integrated area stored in the integrated area feature storage means, the basic area feature stored in the basic area feature storage means, and the combined area feature storage means. A plurality of mathematical formulas for changing each region feature selected from four types of region features or a combination of any of several types of region features among the combined region features and the partial region features calculated by the partial region dividing means are prepared. An appropriate mathematical expression is selected according to the information described in the integrated area design map of the integrated area stored in the integrated area feature storage means, and each feature amount is selected using the selected mathematical expression. The image processing apparatus according to claim 1, wherein the image processing apparatus is changed. For each of the integrated region feature, combined region feature, basic region feature, integrated region, combined region, and basic region that is a description unit of each region feature, the region serial number and the statistics of the pixels included in the region are obtained. It has as an element and contains at least the average color, the barycenter of the pixels included in each area, and the number of pixels included in each area. The image processing apparatus described. The representation means performs a representation process on image data with respect to an integrated area feature stored in the integrated area feature storage means, a basic area feature stored in the basic area feature storage means, or the combined area feature. Data input means is provided for dividing the image data into a plurality of area features so that different processing can be applied by the operator's instructions, with the image range calculated from the combined area features stored in the storage means as a unit. The image processing device according to claim 1, wherein the image processing device is an image processing device. The representation means includes an integrated area feature of the integrated area stored in the integrated area feature storage means, a basic area feature stored in the basic area feature storage means, and a combination stored in the combined area feature storage means. For the region feature, the shape of the brush created from each of the three region features can be variously changed by using the partial region feature calculated from each of the three region features by the partial region dividing means. The image processing apparatus according to any one of claims 1 to 12. The representation means includes an integrated area feature of the integrated area stored in the integrated area feature storage means, a basic area feature stored in the basic area feature storage means, and a combination stored in the combined area feature storage means. Three types of region features, and in addition to these, four types of region features from among the partial region features calculated by the partial region dividing means, or any of several types of region features therein When the position information of the image is used as an input value, a predetermined calculation 3 that minimizes the value when the most preferable region feature is selected is performed, n is a natural number, and the calculation results from the first to the nth smallest value. N basic region features given, or n integrated region integrated region features, or n connected region combined region features, or some of them, and Enter all or some of the difference between the smallest value and the second smallest value (n-1 values calculated from the difference between the smallest value and the (n-1) th smallest value) The image processing apparatus according to claim 1, wherein a pixel feature is calculated as a value. An image processing method for processing an image using a computer,
A divided image data storage step in which the computer stores divided image data obtained as a result of the processing of the basic area dividing processing step of dividing the image data into a plurality of basic areas;
A basic region feature calculating step for calculating a feature for each basic region based on the divided image data stored in the divided image data storage step;
A basic region feature storing step in which the computer stores the basic region feature calculated in the basic region feature calculating step for each region;
An area integration step in which the computer creates an integrated area by integrating a plurality of basic areas based on the basic area features stored in the basic area feature storage step;
An integrated region feature storing step in which the computer stores the integrated region integrated in the region integrating step;
Based on the result calculated from the two area features of the integrated area feature stored in the integrated area feature storage step and the basic area feature stored in the basic area feature storage step, the image data is converted into the image data. An image processing method comprising: an expression step for performing expression processing on the image. An image processing method for processing an image using a computer,
A divided image data storage step for storing divided image data obtained as a result of processing of a basic area dividing processing step in which the computer divides image data into a plurality of basic areas;
A basic region feature calculating step for calculating a feature for each basic region based on the divided image data stored in the divided image data storage step;
A basic region feature storing step for storing the feature calculated by the computer in the basic region feature calculating step for each basic region;
The computer integrates a plurality of basic areas based on the basic area features stored in the basic area feature storage step to create an integrated area, and an area integration step of creating an integrated area design map;
An integrated region feature storing step in which the computer stores the integrated region integrated in the region integrating step;
A region combining step in which the computer combines a plurality of basic regions based on the integrated region feature stored in the integrated region feature storing step to create a combined region;
A combined region feature storing step in which the computer stores the combined regions combined in the region combining step;
All of the integrated region feature stored in the integrated region feature storage step, the basic region feature stored in the basic region feature storage step, and the combined region feature stored in the combined region feature storage step in which the computer stores the combined region, or A partial region dividing step of adaptively dividing a plurality of arbitrary region features into a plurality of portions while referring to the integrated region design map, and calculating a partial region feature for each partial region;
A feature changing step in which the computer changes the integrated region feature, or the basic region feature, or the combined region feature, or the partial region feature;
An expression step in which the computer adaptively expresses the integrated area feature, the basic area feature, the combined area feature, or the partial area feature while referring to the integrated area design map; And an image processing method. An image processing program for processing an image using a computer,
Divided image data storage means for storing divided image data obtained as a result of processing of a basic area dividing processing means for dividing the image data into a plurality of basic areas;
Basic area feature calculating means for calculating a feature for each basic area based on the divided image data stored in the divided image data storage means;
Basic area feature storage means for storing, for each basic area, the basic area feature calculated by the computer in the basic area feature calculating means;
Area integration means for creating an integrated area by integrating a plurality of basic areas based on the basic area features stored in the basic area feature storage means by the computer;
An integrated area feature storage means for storing an integrated area integrated by the computer in the area integration means;
Based on the result calculated from the two area features, the integrated area feature of the integrated area stored in the integrated area feature storage means and the basic area feature stored in the basic area feature storage means, the image data is converted into the image data. An image processing program that causes a computer to function as means for performing expression processing on the computer. An image processing program for processing an image using a computer,
Divided image data storage means for storing divided image data obtained as a result of processing of a basic area dividing processing means for dividing the image data into a plurality of basic areas;
Basic area feature calculating means for calculating a feature for each basic area based on the divided image data stored in the divided image data storage means;
Basic area feature storage means for storing, for each basic area, the feature calculated by the computer in the basic area feature calculating means;
The computer integrates a plurality of basic areas based on the basic area features stored in the basic area feature storage means to create an integrated area, and an area integration means for producing an integrated area design map;
An integrated area feature storage means for storing an integrated area integrated by the computer in the area integration means;
Area combining means for creating a combined area by combining a plurality of basic areas based on the integrated area features stored in the integrated area feature storage means by the computer;
A combined region feature storage means for storing a combined region combined by the computer in the region combining means;
All of the integrated region feature stored in the integrated region feature storage unit and the basic region feature stored in the basic region feature storage unit and the combined region feature stored in the combined region feature storage unit that stores the combined region, or A partial region dividing means for adaptively dividing a plurality of arbitrary region features into a plurality of portions with reference to the integrated region design map and calculating a partial region feature for each partial region, and the computer comprising the integration A feature changing means for changing a region feature, or the basic region feature, or the combined region feature, or the partial region feature;
The computer as means for adaptively expressing the integrated region feature, the basic region feature, the combined region feature, or the partial region feature while referring to the integrated region design map. An image processing program characterized by comprising: In a program that is used in a computer and processes image data,
For each pixel included in the image data, grouping each pixel using at least a color as one of the determination criteria, thereby dividing the image data into a plurality of basic regions;
Calculating a feature of the basic region including at least a color feature for each divided basic region;
For the divided basic regions, a basic region group in which features of the basic regions that are adjacent to each other and are adjacent to each other with respect to the one basic region of interest are smaller than the first threshold is detected, and the basic regions are detected. Similarly, a group of temporary integrated areas is grouped, and a group of basic areas that are also grouped with respect to a basic area other than the one basic area is detected and set as a temporary integrated area. And getting the steps
The first threshold value is made larger than that in the step of obtaining the first temporary integration region, the same steps as the above step are performed, and a plurality of temporary integration regions are obtained, and these are used as the second temporary integration region. The first temporary integration area and the second temporary integration area include a common basic area, and the second temporary integration area corresponds to the number of basic areas included in the first temporary integration area. When the rate of increase in the number of basic areas is smaller than a preset value, determining the provisional integrated area as a formal integrated area;
After the second temporary integration area is determined, the first threshold value is further increased and the third temporary integration area is determined in the same manner. Similarly, the second temporary integration area is formalized. After that, when the same operation is repeated several times, the integration process is continued only for the remaining basic areas excluding the basic area determined as the integrated area in the previous process. Steps,
After the determination process as a formal integration area is completed, for each formal integration area, a step of recording the number of stages indicating how many times the integration process is decided from the temporary integration area to the formal integration area;
An image processing program comprising: calculating, for each of the officially determined integrated areas, characteristics of the integrated area including at least color characteristics based on pixel data included in the basic area. In accordance with the integrated region, determining whether the integrated region is a unit region of expression or a basic region included in the integrated region is a unit region of expression;
24. For each unit area of the expression determined based on the selected result, determining a color of a pixel included in the unit area based on a feature of the unit area of the expression. The image processing program described. The determination as to whether the integrated region is a unit region of expression or a basic region included in the integrated region is a unit region of expression is performed according to the number of stages of the integrated region. 24. The image processing program according to 24.

Images