JP6468897B2 - Image processing apparatus, image processing method, and program - Google Patents

Description

  The present invention particularly relates to an image processing apparatus, an image processing method, and a program suitable for use in dividing an image into a plurality of regions having similar attributes such as coordinates.

  In recent years, a technique for dividing an image into a plurality of regions having similar attributes such as color, texture, and coordinates has attracted attention. The area divided in this way is called a super pixel, and since encoding processing, image processing, image recognition processing, and the like can be executed in units of the divided area, it can be applied to various image processing apparatuses. it can.

  Conventionally, various methods have been proposed for image segmentation. In particular, many techniques for dividing an image into a plurality of regions by clustering pixels have been proposed. In these methods, data having colors, textures, and coordinates as elements are clustered. In the clustering, representative data of each class (region) is obtained, and input data is assigned to the nearest representative data. Here, the nearest representative data is representative candidate data having a shortest distance from the input data among a plurality of representative candidate data for the input data.

  Non-Patent Document 1 proposes a method of dividing into a plurality of regions using a clustering technique called SLIC (Simple Linear Iterative Clustering). SLIC is a technique in which the search range of nearest neighbor representative data in K-means clustering is limited to a local region in an image. In SLIC, a weighted sum of a distance in a coordinate space and a distance in a color space is used as a distance function for clustering. Non-Patent Document 1 proposes ASLIC (Adaptive-SLIC) as a method of adaptively setting the weights of these two distances. In ASLIC, two weights are adaptively set by normalizing the color distance and the coordinate distance using the maximum value of the color space distance and the coordinate space distance from the representative data of each region.

  Further, in the method described in Non-Patent Document 2, in the distance calculation between input data and representative candidate data, a term is introduced in which the distance increases as different classes of input data exist in the vicinity of the input data of interest. . Thus, clustering is performed so that the shape to be divided is close to a circle.

Radhakrishna Achanta, et al., "SLIC Superpixels Compared to State-of-the-Art Superpixel Methods," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 34, No. 11, pp. 2274-2282, Nov. 2012. J. Wang, and X. Wang, "VCells: Simple and Efficient Superpixels Using Edge-Weighted Centroidal Voronoi Tessellations," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 34, No. 6, pp. 1241-1247, Jun. 2012.

  Clustering is a technique often used in image segmentation. However, fragmentation may occur in region division by clustering. Fragmentation is a state in which pixels in the same class are not continuous in the coordinate space. When applying the region division result, it is preferable that no fragmentation occurs.

  The degree of occurrence of fragmentation varies depending on the setting of the color distance weight and the coordinate distance weight. When the weight of the color distance is increased, the tendency of the area to be divided at the edge of the color becomes strong, but fragmentation tends to occur easily. On the other hand, when the weight of the coordinate distance is increased, the tendency tends to be opposite, and fragmentation tends not to occur, but it tends to be difficult to divide the region at the color edge. Thus, there is a trade-off relationship between the region division accuracy and fragmentation, and there is a problem that it is difficult to determine a weight that satisfies the region division accuracy and the degree of fragmentation required by the application destination of region division.

  In the ASLIC disclosed in Non-Patent Document 1, as described above, the color distance and the coordinate distance are normalized using the maximum value of the color distance and the maximum value of the coordinate distance from the representative point data of each region, The weight of the two distances is determined adaptively. However, this method normalizes the coordinate distance and the color distance, and does not determine the weight based on the goal of reducing fragmentation. Therefore, in ASLIC, there is a problem that the result does not necessarily satisfy the region division accuracy and the degree of fragmentation required by the application destination of region division.

  Further, the method described in Non-Patent Document 2 can suppress the occurrence of fragmentation, but it is necessary to refer to the class information of neighboring input data when calculating the distance between each input data and representative candidate data. For this reason, there exists a subject that processing cost is large.

  In view of the above-described problems, an object of the present invention is to suppress the generation of fragmentation while suppressing processing costs when performing image region division with desired region division accuracy.

  The image processing apparatus according to the present invention includes a representative data setting unit that sets representative data for each of a plurality of regions in an image, and the distance between the representative data based on coordinates and colors, respectively. An association means for associating each pixel with any representative data in the plurality of areas, a calculation means for calculating an evaluation value of an area formed by a set of pixels associated by the association means, Target value setting means for setting a target value; weight control means for controlling a weight between the coordinates and the color so that the evaluation value approaches the target value; and the association means includes the weight control Calculating a distance related to the coordinates and color according to the weight controlled by the means, and based on the calculated distance, each pixel in the image is set as any one of the representative data. Characterized in that it put with each other.

  According to the present invention, when image segmentation is performed with desired region segmentation accuracy, it is possible to suppress the processing cost and suppress the occurrence of fragmentation.

It is a flowchart which shows an example of the process sequence of the clustering process in embodiment. It is a figure for demonstrating the block in an image, and its representative data. It is a figure which shows an example of the area division | segmentation result in which fragmentation has generate | occur | produced. It is a block diagram which shows the hardware structural example of the data processor in embodiment.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the drawings. In the present embodiment, an example in which an image is divided into regions using the image region dividing method according to the present invention will be described.

<Configuration example of data processing device>
FIG. 4 is a block diagram illustrating a hardware configuration example of the data processing apparatus 400 that can implement the clustering method according to the present embodiment.
In FIG. 4, a data storage unit 401 is a storage medium that holds image data. The data storage unit 401 includes a hard disk, flexible disk, CD-ROM, CD-R, DVD, memory card, CF card, smart media, SD card, memory stick, xD picture card, USB memory, and the like. In addition to image data, the data storage unit 401 can also store programs and other data. Alternatively, a part of the RAM 406 described later may be used as the data storage unit 401. Furthermore, a storage device of a destination device connected by the communication unit 402 described later may be virtually configured to be used via the communication unit 402.

  The display unit 403 is a device that displays images before and after performing the area division processing, or displays an image such as a GUI, and generally uses a CRT, a liquid crystal display, or the like. Alternatively, it may be an external display device connected by a cable or the like.

  The CPU 404 executes main processing according to the present embodiment and controls the overall operation of the data processing device 400. The ROM 405 and the RAM 406 provide the CPU 404 with programs, data, work areas, and the like necessary for the processing. When a program necessary for processing to be described later is stored in the data storage unit 401 or stored in the ROM 405, the processing described later is executed after being once written in the RAM 406. When the data processing apparatus 400 receives a program via the communication unit 402, the program is temporarily recorded in the data storage unit 401 and then written in the RAM 406 or directly from the communication unit 402 to the RAM 406. Executed.

  The CPU 404 reads image data written in the RAM 406 or the data storage unit 401 and executes image processing (clustering processing) according to the present embodiment. Data in the middle of processing such as representative data is written into the RAM 406 by the CPU 404 and is read out when necessary. Then, the clustering processing result is written in the RAM 406. Further, the CPU 404 performs control for displaying an image on the display unit 403, and performs control for transmitting image data to an external apparatus via the communication unit 402.

  The communication unit 402 is an I / F for performing communication between devices. As the I / F, for example, a well-known local area network, USB, IEEE1284, IEEE1394, a wired communication system such as a telephone line may be used. Further, a wireless communication system such as infrared (IrDA), IEEE 802.11a, IEEE 802.11b, IEEE 802.11g, IEEE 802.11n, Bluetooth (registered trademark), UWB (Ultra wide Band), or the like may be used.

  In the example shown in FIG. 4, an example is shown in which the data storage unit 401 and the display unit 403 are all included in one device, but these parts are connected by a communication path using a known communication method. Therefore, the configuration as a whole may be used. Further, in the example shown in FIG. 4, the configuration has only one CPU, but a configuration in which a plurality of CPUs are provided may be employed. As for the system configuration, various components other than those described above may exist, but the description thereof is omitted because it is not the main point of the present invention.

  Hereinafter, an outline of processing performed by the data processing apparatus 400 illustrated in FIG. 4 will be described. In this embodiment, pixel data is used as two-dimensional coordinates, three-dimensional color features, and five-dimensional input data in total, and this input data is clustered into representative data to perform region division. At this time, the distance between the representative data and the input data is defined by a weighted sum of a distance in the coordinate space (hereinafter referred to as coordinate distance) and a distance in the color space (hereinafter referred to as color distance).

  Since the weights of the coordinate distance and the color distance are parameters that influence the result of area division, it is necessary to set appropriate weights in order to obtain the area division accuracy and area shape necessary for the application of area division. Therefore, in the present embodiment, an evaluation function for evaluating the region division result and its target value are defined, and the weight is controlled so that the evaluation value of the region division result approaches the target value. In this embodiment, in order to reduce fragmentation of a region, the complexity of the region shape is used as an evaluation function, and the weight is controlled in units of regions so that the complexity of each region approaches the target value.

<Clustering for region division>
FIG. 1 is a flowchart illustrating an example of a processing procedure of clustering processing in the present embodiment. Each step shown in FIG. 1 indicates the operation of the CPU 404.
First, clustering target data will be described. Equation (1) below shows the i-th input data. The input data is a five-dimensional vector corresponding to the pixel in the image, and the i-th input data is information regarding the i-th pixel read out by raster scanning the image.
X _i = (x _{i, 1} , x _{i, 2} , x _{i, 3} , x _{i, 4} , x _{i, 5} ) (1)

In formula (1), x _{i, 1} and x _{i, 2} represent positions in a two-dimensional coordinate space, and mean coordinate values in the vertical and horizontal directions of the image, respectively. X _{i, 3} , x _{i, 4} and x _{i, 5} are three-dimensional color features representing the color components of the pixels.

First, in step S101, initialized representative data is generated. Specifically, as representative data setting processing, for example, the image 201 shown in FIG. 2 is divided into blocks on the coordinate space, and one representative data 205 is set in each block 203. Then, input data in the block 203 is selected at random, and a vector having the coordinates and color features of the selected input data as elements is set as a vector of the representative data. Equation (2) below shows the j-th representative data C _j . Here, j is an index assigned to the representative data arranged as shown in FIG. 2 in raster order.
C _j = (c _{j, 1} , c _{j, 2} , c _{j, 3} , c _{j, 4} , c _{j, 5} ) (2)

In equation (2), c _{j, 1} and c _{j, 2} represent positions in a two-dimensional coordinate space, and mean coordinate values in the vertical and horizontal directions of the image, respectively. Further, c _{j, 3} , c _{j, 4} , and c _{j, 5} are three-dimensional color features representing the color components of the pixels.

  In step S102, the target value of the evaluation value of the region division result is set. The definition of the evaluation function of the area division result and the target value will be described later.

In step S103, a predetermined maximum number of repetitions is set, the number of repetitions t is controlled, and a repetition calculation loop from steps S104 to S113 is started. The initial value of t is 1.
In step S104, a loop of input data is started. The order i of the input data X _i in the image is controlled, and the processing from step S105 to step S109 is repeated to cluster all input data.

  First, in step S105, a search range is set. Specifically, the search range for representative candidate data is set according to the coordinate position of the input data. Here, the representative candidate data of the input data is a candidate of representative data to which the corresponding input data is assigned (belongs to).

  Here, an example of representative data in which the width and height of the search range 202 are each five will be described. First, a search range of representative candidate data is set around a block including input data to be processed (hereinafter referred to as attention input data) 206. In this case, the search range is representative data in the vicinity of 5 × 5 coordinates. The search range 202 has a vertical direction of 5 to 9, and a horizontal direction of 6 to 10. In the case of the example shown in FIG. 2, since the number of representative data per line is 15, the indexes included in the search range 202 are 66 to 70, 81 to 85, 96 to 100, 111 to 115, 126 to 130. It is. Therefore, the search range 202 of all input data included in the range of the block 204 including the target input data 206 is the same. It is assumed that the relationship between the position of the input data and the representative candidate data in the search range is fixed first, and the index of the representative data does not change even if the position of the representative data is changed by iterative processing.

Steps S106 to S108 are steps for searching the nearest representative data of the input data. Here, the nearest representative data is representative candidate data having the shortest distance from the input data, that is, the nearest representative data in the distance space. Before describing the processing contents in each step, a method for calculating the distance between input data and representative candidate data in the present embodiment will be described. The following equation (3) is a distance function representing a distance D (X _i , C _j ) between the i-th input data and the j-th representative candidate data.
D (X _i , C _j ) = W (X _i , t) × D _s (X _i , C _j ) + D _c (X _i , C _j ) (3)

In the formula (3), D _s represents a coordinate distance, and D _c represents a color distance. W (X _i , t) is a weight of the coordinate distance D _s in the t-th iteration of the i-th input data X _i . As shown in the following formula (4), the weight W (X _i , t) is the weight W _{j, t} corresponding to the region S _{j, t−1} to which the input data X _i belonged in the previous iteration. Become. However, in the first iteration, the initial value W ₁ is used as the weight of the coordinate distance for all X _i . In the present embodiment, the initial value W ₁ is 1.

  When the weight of the coordinate distance is increased, the area is divided with more importance on the coordinate distance than the color distance, so even if the color is slightly different between the representative data and the input data, the input of the coordinate position from the representative data is close. Data tends to be divided into the same areas. On the other hand, if the weight of the coordinate distance is reduced, the area is divided by placing more emphasis on the color distance than the coordinate distance, so even if the coordinate position from the representative data is slightly separated, the input is similar in color to the representative data. Data tends to be divided into the same areas. Therefore, the fragmentation is reduced as the weight of the coordinate distance is increased, but the region is more difficult to be divided at the boundary of the object. On the other hand, the smaller the weight of the coordinate distance, the easier the region is divided at the boundary of the object, but the fragmentation tends to increase.

As shown in equation (3), the distance D between the input data and the representative candidate data is calculated by the weighted sum of the coordinate distance D _s and the color distance D _c. Hereinafter, such a distance calculated from a plurality of partial distances is referred to as a total distance. The following formulas (5) and (6) are calculation formulas for calculating the coordinate distance D _s and the color distance D _c , respectively. In the present embodiment, the Euclidean distance is used, but the present invention is not limited to this.

In step S106, a representative candidate data loop is started. The order j of the representative candidate data C _j is controlled to search for all representative candidate data. In step S107, the total distance between the i-th input data and the j-th representative candidate data is calculated according to equation (3). In step S108, it is determined whether or not the calculation of the total distance with all representative candidate data has been completed. As a result of the determination, if the total distance with all the representative candidate data is calculated, the process proceeds to step S109 to end the representative candidate data loop. On the other hand, if there is representative candidate data for which the total distance has not yet been calculated, the process returns to step S107 to calculate the total distance with the next representative candidate data.

  Subsequently, in step S109, the input data is attributed (associated) to the representative candidate data (nearest neighbor representative data) having the shortest distance among the total distances calculated in the loop of steps S106 to S108. In step S110, it is determined whether all input data have been processed. If all the input data has been processed as a result of this determination, the process proceeds to step S111, and the input data loop is terminated. On the other hand, if there is input data that has not been processed yet, the process returns to step S104 to start processing the next input data.

In step S111, an average vector of the assigned input data vectors is calculated for each representative data. Then, the average vector is set as a new vector C _j of the corresponding representative data, and the representative data is updated. Next, in step S112, the updated evaluation value of each area is calculated. A method for calculating the evaluation value will be described later.

Subsequently, in step S113, the weight W (X _i , t) used when calculating the total distance between the input data and the representative candidate data in step S108 is controlled. In this process, the weight is controlled so that the evaluation value calculated in step S112 approaches the target value set in step S102. This weight control will be described later.

  Next, in step S114, it is determined whether or not the maximum number of repetitions has been exceeded, or whether or not the clustering result has converged. If any of the conditions is satisfied as a result of this determination, the iterative calculation is terminated. On the other hand, if none of the conditions is satisfied, the process returns to step S104 to start the next iterative calculation. When the above processing is completed, the CPU 404 outputs the clustering result (region division result) to the display unit 403 or the like.

<Evaluation value calculation of area>
Next, calculation of the evaluation value in step S112 will be described. In the present embodiment, the complexity of the area shape is used as the evaluation value for each area. The evaluation value of each area is used for the weight control of the coordinate distance in step S113. Each area is constituted by a set of input data belonging to the representative data. The following formulas (7) to (11) show evaluation value calculation formulas.

Area S _j, evaluation value of _t E (S _{j, t),} as shown in Equation (7), region S _j, the area of _t E _s (S _{j, t)} and, region S _j, the contour of _t It is calculated using the length E _c (S _{j, t} ). The area E _s (S _{j, t} ) is defined by the number of elements in the region S _{j, t} as shown in Expression (8). The contour length E _c (S _{j, t} ) is the total number of input data determined as a region boundary among the input data X _i that is an element of the region S _{j, t} as shown in the equation (9).

“Contour (X _i )” is a function for determining whether or not the input data X _i is a region boundary. As shown in the equation (10), the region number to which the input data in the vicinity of the input data X _i around 8 belongs. Calculated with reference to If all the input data in the vicinity of the surrounding 8 belong to the same region as the target input data X _i , the target input data is determined to be a non-region boundary, and Contour (X _i ) becomes zero. If there is at least one input data belonging to a region different from the target input data X _{i in the} vicinity of the surrounding eight input data, the target input data is determined to be a region boundary, and Control (X _i ) is 1. It becomes. L (x, y) represents a region number to which the input data at the coordinate position (x, y) belongs, and δ (x) is a function that is 1 when x = 0 and 0 for other x.

The evaluation value E (S _{j, t} ) indicates the complexity of the area. For example, as shown in FIG. 3, fragmentation occurs in the region 301 to the region 303, the region 304 to the region 307, the region 308 to the region 315, and the region 316 to the region 317. As in these regions, when the unevenness is large, or when the regions S _{j, t} are not continuous in the coordinate space, the complexity increases. When all input data included in the region S _{j, t} are discontinuous in the coordinate space, the contour length of the region is equal to the area, and the complexity is 1. On the other hand, when there are few unevenness | corrugations like the area | region 318 and all the input data are adjacent on coordinate space, complexity becomes small. As described above, an area where fragmentation occurs tends to increase in complexity, and an area where fragmentation does not occur tends to decrease in complexity.

<Set weight target value>
In step S102, the target value of the evaluation value calculated in step S112 is set. In the present embodiment, the target value T of the evaluation value E (S _{j, t} ) is calculated by the following equation (12). Target value T is as shown in equation (12), the evaluation value when the area shape becomes horizontal L _x, a rectangular shape of vertical L _y, is a value in consideration of the 20% of the complexity. The lengths of the horizontal L _x and the vertical L _y are N _x images of width I _x and height I _{y in} the horizontal direction and N _{y in} the vertical direction, as shown in equations (13) and (14), respectively. The length of the side when dividing into the uniform rectangular area is shown.

<Weight control>
In step S113, using the target value T set in step S102 and the evaluation value E (S _{j, t} ) calculated in step S112, the weight W _{j, t + 1} of the coordinate distance in the (t + 1) th iteration is calculated. calculate. The following equation (15) shows a weight control equation. k is a control coefficient indicating the strength of control. In this embodiment, the value of k is set to 0.1. In the first th iteration and controls the initial value W ₁ of the weight, in the second and subsequent repetitive performs control for the current weight W _{j, t.}

  With the control of Expression (15), the weight of the coordinate distance is increased in the region where the shape of the region is more complex than the target, so that the region shape becomes simple in the next region division, and the evaluation value is expected to be small. it can. In the region where the shape of the region is simpler than the target, the weight of the coordinate distance decreases, so that the region shape becomes complicated and the evaluation value can be expected to increase in the next iteration.

  As described above, according to the method of the present embodiment, since the weight of the coordinate distance is controlled so that the complexity of the shape of each region approaches the target value, fragmentation can be reduced. .

(Other embodiments)
In the embodiment described above, an example in which the weight of the coordinate distance is controlled in units of regions has been described, but the present invention is not limited to this, and the weight can be controlled in units of blocks of the image. Equation (16) below shows the weight W (X _i , t) of the input data of interest X _i .

B _j in equation (16) is the block shown in FIG. 2 and is a set of input data. For example, the block including the attention input data 206 is the block 204. As shown in FIG. 2, each block has corresponding representative data. Therefore, the weight W _{j, t} corresponding to the j-th block B _j is controlled using the j-th region S _{j, t} as shown in Expression (15). When the weight is controlled in units of regions as in the above-described embodiment, it is necessary to refer to the region number to which the input data X _i belonged in the previous iteration. On the other hand, when the weight is controlled in units of blocks, there is an advantage that the processing cost is smaller because it is not necessary to refer to the area number.

It is also possible to set a weight for each image using the following equation (17). W _t is a weight in the t-th iteration. W _t can be obtained by using the evaluation value average E (t) of each region as shown in the following formula (18). The evaluation value average E (t) is calculated by the following equation (19). When the weight is controlled in units of images, there is an advantage that the processing cost is lower than that in the case of controlling the weights in units of blocks because the weight to be controlled is one.

  In the above-described embodiment, the complexity of the region shape shown in Expression (7) to Expression (11) is used as the evaluation function, but the present invention is not limited to this. An arbitrary function for evaluating the region division result can be applied to the evaluation function. For example, it is also possible to divide the region so that the variance value of the color feature of each region is approximately the same, using a variance value that is a statistic of the color feature of the region as the evaluation value.

  In the above-described embodiment, one weight is controlled by one evaluation function and target value, but the present invention is not limited to this. It is also possible to control a plurality of weights by a plurality of evaluation functions and target values.

  In the above-described embodiment, Equations (12) to (14) are used as the method for calculating the target value of the evaluation value of each region, but the present invention is not limited to this. As the target value, an arbitrary value can be set from a range that the evaluation function can take.

In addition, since the region division result by the processing of the above-described embodiment depends on the initial value W _{1 of the} weight, a good region division result may not be obtained depending on the setting of the initial value. Therefore, it is possible to perform the processing of steps S101 to S114 in FIG. 1 a plurality of times while changing the initial value of the weight, and select the best region division result as the final output.

  In the embodiment described above, a constant value is set for the control coefficient k when controlling the weight, but the present invention is not limited to this. The control coefficient k may be defined as a monotonically decreasing function k (t) with a repetition number t as shown in the following equation (20). Note that e indicates the base of the natural logarithm. Since equation (20) is a monotonously decreasing function depending on the number of repetitions t, the weights converge without oscillating during the repetition of step S103.

  In order to reduce the influence of the initial value of the weight, the disturbance term R (t) can be added to the weight control expression, and the weight can be controlled using the following expressions (21) and (22). . Note that rand (−1, 1) is a random number in the range [−1, 1], and the disturbance term R (t) is a term in which the range width monotonously decreases with the number of repetitions t. When Expression (21) and Expression (22) are used, during the beginning of the repetition of Step S103, the weight changes due to strong disturbance that is not affected by the evaluation value and the target value. It can be reduced. Furthermore, when a monotonically decreasing function such as Equation (20) is used for the control coefficient k, the weight can be converged stably by setting the convergence speed of the disturbance term larger than the convergence speed of the control coefficient k. Is possible.

  The present invention supplies a program that realizes one or more functions of the above-described embodiments to a system or apparatus via a network or a storage medium, and one or more processors in a computer of the system or apparatus read and execute the program This process can be realized. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

401 Data storage unit 404 CPU

Claims (10)

Representative data setting means for setting representative data for each of a plurality of regions in the image;
An associating means for associating each pixel in the image with any representative data in the plurality of regions based on a coordinate and color distance with the representative data;
Calculating means for calculating an evaluation value of an area composed of a set of pixels associated by the associating means;
Target value setting means for setting a target value of the evaluation value;
Weight control means for controlling the weight between the coordinates and the color so that the evaluation value approaches the target value;
The associating means calculates a distance related to the coordinates and color according to the weight controlled by the weight control means, and based on the calculated distance, each pixel in the image is set as any one of the representative data. An image processing apparatus characterized by associating.   The image processing apparatus according to claim 1, wherein the representative data setting unit updates the respective representative data based on the associated set of pixels.   3. The image processing apparatus according to claim 1, wherein the weight control unit controls a weight between the coordinates and the color in units of an area formed by the set of associated pixels. 4. .   The image processing apparatus according to claim 1, wherein the weight control unit controls a weight between the coordinates and the color in units of blocks.   The image processing apparatus according to claim 1, wherein the weight control unit controls a weight between the coordinates and the color in units of images.   The image processing apparatus according to claim 1, wherein the calculation unit calculates, as the evaluation value, a complexity of a shape of an area configured by the associated set of pixels. .   The image processing according to claim 1, wherein the calculation unit calculates a statistic of a color feature in an area configured by the set of associated pixels as the evaluation value. apparatus.   The image processing according to any one of claims 1 to 7, further comprising output means for outputting a result of an area formed by the associated set of pixels when a predetermined condition is satisfied. apparatus. A representative data setting step for setting representative data for each of a plurality of regions in the image;
An association step of associating each pixel in the image with any representative data in the plurality of regions based on a coordinate and color distance with the representative data;
A calculation step of calculating an evaluation value of an area configured by a set of pixels associated in the association step;
A target value setting step for setting a target value of the evaluation value;
A weight control step of controlling a weight between the coordinates and the color so that the evaluation value approaches the target value,
In the association step, a distance related to the coordinates and color is calculated according to the weight controlled in the weight control step, and each pixel in the image is set to any one of the representative data based on the calculated distance. An image processing method characterized by associating with. A representative data setting step for setting representative data for each of a plurality of regions in the image;
An association step of associating each pixel in the image with any representative data in the plurality of regions based on a coordinate and color distance with the representative data;
A calculation step of calculating an evaluation value of an area configured by a set of pixels associated in the association step;
A target value setting step for setting a target value of the evaluation value;
Causing a computer to execute a weight control step of controlling a weight between the coordinates and the color so that the evaluation value approaches the target value;
In the association step, a distance related to the coordinates and color is calculated according to the weight controlled in the weight control step, and each pixel in the image is set to any one of the representative data based on the calculated distance. A program characterized by associating with.

Images