JP2009225211A - Information processing apparatus and method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To more accurately detect an iterative pattern in an image. <P>SOLUTION: An image holding unit 101 is constituted of a storage medium such as, e.g., a semiconductor memory, and holds an input image input to a pattern determination device 100. An image evaluation processing unit 102 evaluates an input image being held in the image holding unit 101 according to a predetermined method. An iterative pattern detection unit 103 uses evaluation values being held in an evaluation value holding unit 117 to perform iterative pattern detection processing, determines whether an iterative pattern is included in the input image, and outputs the result of the determination to the outside of the pattern determination device 100. The present invention can be applied to, e.g., pattern determination devices. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an information processing device and method, and a program, and more particularly, to an information processing device and method, and a program that can detect a repeated pattern in an image more accurately.

  Conventionally, it has been considered to analyze the contents of an image and use the detection result for image processing or the like. For example, in a moving image such as a video image taken by a video camera, a desired portion of each frame image is specified as a tracking point, and the image is enlarged or the video camera operation is controlled so as to track the tracking point. (For example, refer to Patent Document 1). It is also conceivable to detect an edge from the image and specify a face or person image, or to perform special image processing on the edge portion.

  In particular, a characteristic pattern included in an image, for example, a repeated pattern in which a predetermined image pattern such as a straight line or a geometric pattern is repeated can be used for various types of processing. It is useful to detect such a repeated pattern. It is.

  Various methods are conceivable as a method for detecting such a repeated pattern (see, for example, Patent Documents 2 to 4).

JP 2005-303983 A JP 2005-56410 A JP 2006-133752 A JP 2004-264924 A

  However, the methods described in Patent Document 2 and Patent Document 3 detect the periodicity of an image, and can only detect an image of a repetitive pattern having periodicity, such as a straight edge. There is a fear that it is difficult to detect a repetitive pattern having no periodicity.

  Further, in Patent Document 4, it is possible to detect a straight line by block matching. However, in this method, since it is selected from several kinds of predefined straight lines, it is difficult to detect a straight line in an arbitrary direction. There was a fear.

  The present invention is for solving such a problem, and makes it possible to detect a repeated pattern in an image more accurately.

  One aspect of the present invention is an information processing apparatus that detects a repetitive pattern in which a predetermined image pattern included in an image is repeated, and is a predetermined region corresponding to a point of interest that is a pixel to be processed of the image Pixels other than the target point for evaluating each pixel in the target block and analyzing the distribution of the evaluation value in the target block in the target block for evaluating each pixel Feature point specifying means for specifying a plurality of feature points, and determination means for determining whether or not the image includes the repetitive pattern based on a relationship between evaluation values of the specified feature points. An information processing apparatus is provided.

  The evaluation value is obtained by comparing the target block with a target block that is a predetermined area corresponding to a comparison point that is a pixel in the target block selected as a target for comparison. It can be a parameter indicating the similarity to the block of interest.

  The evaluation value of the comparison point may be a sum of absolute values of difference values of pixels having the same position in the block of the comparison target block corresponding to the comparison point and the block of interest. it can.

  The feature point specifying means can specify three points as the feature points.

  The repetitive pattern is an image of a linear edge, and the determination unit determines whether or not there is an image of the linear edge, and when determining that the image of the linear edge exists, The direction can also be determined.

  The feature point specifying means specifies, as a first feature point, a comparison point of an evaluation value indicating that the similarity of the corresponding comparison target block to the target block is the highest except for the target point, and Two points on the straight line orthogonal to the straight line passing through the first feature point and the point of interest, which are separated by a predetermined distance from the point of interest, are specified as the second feature point and the third feature point. Can do.

  The feature point specifying means can specify the first to third feature points from pixels in a region excluding a range within a predetermined distance from the target point in the target block.

  The apparatus further comprises feature analysis means for analyzing a feature of the evaluation value distribution by calculating an average value of the evaluation values of the first feature point and the second feature point, and the determination means includes the feature analysis The average value calculated by the means and the evaluation value of the first feature point are compared, and if the difference is larger than a predetermined threshold, the straight line passing through the first feature point and the point of interest It can be determined that a straight edge exists.

  Based on the determination result by the determination unit, the image determination unit further includes a re-determination unit that determines again whether the image includes the repetitive pattern, and the image evaluation unit sets the evaluation value for each pixel of a plurality of blocks of interest. The feature point specifying unit specifies a plurality of feature points in each of the plurality of blocks of interest, and the determination unit determines whether the image includes the repetitive pattern for each of the plurality of blocks of interest. The re-determination unit can determine again whether or not the image includes the repetitive pattern based on a plurality of determination results obtained by the determination by the determination unit.

  Another aspect of the present invention is an information processing method for an information processing apparatus that detects a repetitive pattern in which a predetermined image pattern is included, which is included in an image, and corresponds to an attention point that is a pixel to be processed in the image Other than the point of interest for evaluating each pixel in the target block, which is a predetermined area to be obtained, obtaining each evaluation value, and analyzing the distribution of the evaluation value in the target block in the target block An information processing method including a step of identifying a plurality of feature points that are pixels and determining whether the image includes the repetitive pattern based on a relationship between evaluation values of the plurality of identified feature points. is there.

  According to another aspect of the present invention, in a program for detecting a repetitive pattern in which a predetermined image pattern included in an image is repeated, a target block that is a predetermined region corresponding to a target point that is a pixel to be processed of the image An image evaluation step for evaluating each pixel in the image and obtaining an evaluation value thereof, and a pixel other than the target point for analyzing the distribution of the evaluation value in the target block in the target block A feature point specifying step for specifying a plurality of feature points and a determination step for determining whether or not the image includes the repetitive pattern based on the relationship between the evaluation values of the plurality of specified feature points. It is a program to be executed.

  In one aspect of the present invention, each pixel in a target block that is a predetermined area corresponding to a target point that is a pixel to be processed of an image is evaluated, and each evaluation value is obtained. A plurality of feature points that are pixels other than the point of interest for analyzing the distribution of evaluation values in the block of interest are identified, and the image repeats based on the relationship between the evaluation values of the plurality of identified feature points. Whether or not is included.

  According to the present invention, information can be processed. In particular, the repeated pattern in the image can be detected more accurately.

  FIG. 1 is a block diagram showing a main configuration example of a pattern determination apparatus to which the present invention is applied. A pattern determination apparatus 100 shown in FIG. 1 is an apparatus that detects a repetitive pattern in which a predetermined image pattern such as a straight line or a geometric pattern is included in an image of input image data. As illustrated in FIG. 1, the pattern determination apparatus 100 includes an image holding unit 101, an image evaluation processing unit 102, and a repeated pattern detection unit 103.

  The image holding unit 101 is configured by a storage medium such as a semiconductor memory, for example, and holds an input image input to the pattern determination apparatus 100.

  The image evaluation processing unit 102 evaluates one input image held in the image holding unit 101 by a predetermined method. This evaluation method is arbitrary as long as the repeated pattern detection unit 103 can detect a repeated pattern included in the input image. For example, the image evaluation processing unit 102 evaluates the input image using a block matching process that compares blocks included in the input image.

  The block is a data unit that divides the entire input image by a predetermined image size such as 32 pixels × 32 pixels. That is, the block is a partial image that constitutes a partial region of the entire input image. The block matching process is a process for comparing two blocks and calculating an evaluation value representing the similarity between the images.

  In general, there are many block matching methods, but basically any method can be used as long as it can evaluate an image. However, general block matching is often performed between two consecutive frame images of a moving image, but the image evaluation processing unit 102 extracts different regions (positions) extracted from one input image. Compare two blocks.

  In general, block matching processing is a technique widely used in image processing. Therefore, by using such general block matching processing for image evaluation, development is easier than evaluation using an original evaluation method. Become. Further, for example, when block matching processing is performed in other image processing in the apparatus, the algorithm can be used. Note that, for example, in a case where a similar block matching process is performed in a process performed in another apparatus or another processing unit such as an image process performed using a determination result by the pattern determination apparatus 100, the image evaluation processing unit 102 However, it is also possible to use the processing result. Thus, the development cost of the image evaluation processing unit 102 can be reduced by using the block matching process.

  In the following, an example in which block matching processing is performed as image evaluation processing will be described. FIG. 1 shows an example of a detailed configuration of the image evaluation processing unit 102 that performs block matching processing. In FIG. 1, an image evaluation processing unit 102 includes an attention point specifying unit 111, an attention block specifying unit 112, a comparison point selecting unit 113, a comparison target block specifying unit 114, a block matching unit 115, an evaluation value calculating unit 116, and an evaluation value. A holding portion 117 is included.

  The attention point identification unit 111 identifies the attention point from the images held in the image holding unit 101. The attention point indicates a pixel (processing target pixel) for detecting a repetitive pattern in the input image. The attention block identification unit 112 identifies the attention block from the same image. The target block indicates a block on which block matching is performed for detecting a repetitive pattern at the target point, and indicates a region including the target point. Note that the size and shape of the block of interest are arbitrary. An attention point is a reference point for the attention block, and there is one attention point for the attention block. Although the point of interest may be located at any position of the block of interest, the following description will be made assuming that the point of interest is the center point of the block of interest.

  Note that the attention point specifying unit 111 and the attention block specifying unit 112 set information on the attention point and the attention block, for example, which position of the input image is the attention point, and how to set the attention block with respect to the attention point. How to obtain such information is arbitrary. For example, the user may input to the pattern determination apparatus 100 using an input unit (not shown), and the attention point specifying unit 111 and the attention block specifying unit 112 may use the information. Further, for example, the attention point specifying unit 111 and the attention block specifying unit 112 may acquire and use information supplied from an image processing device or the like subsequent to the pattern determination device 100. Of course, the attention point specifying unit 111 and the attention block specifying unit 112 may hold these pieces of information in advance.

  The comparison point selection unit 113 specifies a comparison point from the images held in the image holding unit 101. The comparison target block specifying unit 114 specifies a comparison target block from the same image. The comparison point is a reference point for a comparison target block that is a block that performs block matching with the target block, and one comparison point is associated with one comparison target block. The size and shape of the comparison target block are the same as the target block, and the position of the comparison point in the comparison target block is the same as the position of the target point in the target block.

  For example, the block matching unit 115 calculates a difference between each pixel value of the block of interest and each pixel value of the comparison target block between pixels having the same position in the block (hereinafter referred to as an intra-block position). Thus, the target block and the comparison target block are compared. Note that this block-to-block comparison method is arbitrary, and comparison may be made by other methods.

  The evaluation value calculation unit 116 calculates an evaluation value indicating the similarity between the target block and the comparison target block based on the comparison result between the target block and the comparison target block by the block matching unit 115. For example, the evaluation value calculation unit 116 calculates, as the evaluation value, the sum of the absolute values of the difference values between the pixels of the same block in the target block and the comparison target block calculated by the block matching unit 115. Note that this similarity evaluation method is arbitrary. In addition to this, for example, the average value of the absolute values of the difference values of the pixels having the same in-block position between the target block and the comparison target block is used as the evaluation value. The block matching unit 115 may select some pixels of the target block and the comparison target block as representative points, obtain a difference value between the representative points, and the evaluation value calculation unit 116 may The evaluation value may be calculated using only the obtained difference value.

  The evaluation value holding unit 117 is configured by a storage medium such as a semiconductor memory, for example, and holds the evaluation value calculated in this way for each comparison target point.

  Note that the comparison point selection unit 113 sequentially selects all or some of the predetermined pixels in the target block as comparison points sequentially. That is, a comparison target block is set for each comparison point, and block matching processing is performed between the block of interest and each comparison target block. The evaluation value holding unit 117 sequentially holds the block matching processing results.

  The repetitive pattern detection unit 103 performs repetitive pattern detection processing using those evaluation values held in the evaluation value holding unit 117, determines whether or not a repetitive pattern is included in the input image, and determines the determination result. Output to the outside of the pattern determination apparatus 100. This determination result is used for image processing or the like performed in the subsequent stage. As shown in FIG. 1, the repetitive pattern detection unit 103 includes a feature point specifying unit 121, a feature analysis unit 122, and an analysis result determination unit 123.

  The feature point specifying unit 121 specifies pixels other than the target point indicating the feature of the target block as the feature point based on the evaluation value group corresponding to each pixel of the target block held in the evaluation value holding unit 117. For example, the feature point specifying unit 121 specifies the comparison point having the highest similarity (of the comparison target block) as the first feature point, and is perpendicular to the first straight line connecting the first feature point and the point of interest. Two points in the target block that are on the second straight line passing through the target point and are separated from the target point in a direction opposite to each other by a predetermined distance are specified as the second feature point and the third feature point. At this time, for example, the feature point specifying unit 121 specifies a feature point in a region that is in the target block and that is separated from the target point by a predetermined distance or more. Note that the method for specifying feature points is arbitrary, and methods other than those described above may be used.

  The feature analysis unit 122 analyzes the feature of the block of interest using the feature points specified by the feature point specification unit 121. For example, the feature analysis unit 122 calculates an average value of the evaluation values of the second feature point and the third feature point. It is determined whether the difference between the average value and the first feature point is greater than a predetermined threshold. This analysis method is arbitrary, and methods other than those described above may be used.

  The analysis result determination unit 123 determines whether or not the block of interest includes a repetitive pattern based on the analysis result of the feature analysis unit 122. This determination method is arbitrary. For example, when the average value of the evaluation values of the second feature point and the third feature point and the difference value of the first feature point are larger than a predetermined threshold value, a repetitive pattern is included. If the difference value is smaller than the threshold value, it is determined that the repeated pattern is not included.

  Next, an example of the flow of pattern determination processing executed by the pattern determination apparatus 100 will be described with reference to the flowchart of FIG.

  When an image is input to the pattern determination apparatus 100, the pattern determination apparatus 100 starts pattern determination processing. When the pattern determination process is started, the image holding unit 101 holds the image data of the input image in step S1. When at least one input image is held, the image evaluation processing unit 102 performs image evaluation processing using the input image held in step S1 in step S2. When the image evaluation process ends, the repeated pattern detection unit 103 detects a repeated pattern from the input image using the image evaluation process result in step S3, and outputs the detection result.

  As described above, the pattern determination apparatus 100 performs the image evaluation process, performs the repeated pattern detection process using the evaluation result, and detects the repeated pattern from the input image.

  Next, an example of a detailed flow of the image evaluation process executed in step S2 of FIG. 2 will be described with reference to the flowchart of FIG.

  When the image evaluation process is started, in step S21, the point-of-interest specifying unit 111 specifies the point of interest from the input image based on information given in advance from, for example, a user or a subsequent processing unit.

  When the attention point is specified, the attention block specifying unit 112 determines the reference point specified by the process of step S21 as a reference point in step S22 based on, for example, information given in advance by the user, a subsequent processing unit, or the like. A target block of a predetermined size is specified from the input image.

  When the block of interest is specified, the comparison point selection unit 113 selects pixels (unprocessed pixels) that have not been set as comparison points in the block of interest in a predetermined order in step S23, and selects the pixels as comparison points. And In step S24, the comparison target block specifying unit 114 specifies a comparison target block having a predetermined size with the comparison point as a reference point. Note that the shape and size of this comparison target block are the same as the target block identified in step S22.

  When the target block and the comparison target block are specified, the block matching unit 115 compares the target block with the comparison target block in step S25. For example, the block matching unit 115 compares each pixel value of the block of interest with each pixel value of the comparison target block between pixels in the same block, and obtains a difference. Then, in step S26, the evaluation value calculation unit 116 calculates the evaluation value of the comparison point based on the comparison result obtained by the process of step S25. For example, the evaluation value calculation unit 116 calculates the sum of the absolute values of the difference values between the pixels in the same block of each pixel value of the target block and each pixel value of the comparison target block as the evaluation value.

  In step S27, the evaluation value holding unit 117 stores the evaluation value calculated by the process of step S26 in association with the comparison point. In step S28, the comparison point selection unit 113 determines whether or not there is an unprocessed pixel that is not selected as a comparison point in the target block. If it is determined that there is an existing block, the process returns to step S23, and thereafter Repeat the process. That is, the comparison point selection unit 113 to the evaluation value holding unit 117 repeat the processing from step S23 to step S28 until all the pixels in the block of interest are selected as comparison points. If it is determined in step S28 that there is no unprocessed pixel, the comparison point selection unit 113 ends the block matching process, returns the process to step S2 in FIG. 2, and executes the processes after step S3.

  Next, an example of a detailed flow of the repeated pattern detection process executed in step S3 of FIG. 2 will be described with reference to the flowchart of FIG.

  When the repeated pattern detection process is started, the feature point specifying unit 121 sets the feature points in the target block to pixels other than the target point based on the evaluation value held by the evaluation value holding unit 117 in step S41. Specify more than one. In step S42, the feature analysis unit 122 analyzes the feature of the evaluation value at the feature point specified by the processing in step S41, and generates an analysis result indicating the presence or absence of the repeated pattern. In step S43, the analysis result determination unit 123 determines whether the image of the block of interest (input image) includes a repetitive pattern based on the analysis result obtained by the process of step S42.

  When the analysis result determination ends, the analysis result determination unit 123 ends the repeated pattern detection process, returns the process to step S3 in FIG. 2, and ends the pattern determination process.

  As described above, by evaluating the input image, identifying and analyzing the feature points based on the evaluation result, and detecting the repeated pattern, the pattern determination apparatus 100 can more accurately repeat the repeated pattern in the image. Can be detected.

  A specific example of repeated pattern detection will be described with reference to FIGS.

  A partial image 201 in FIG. 5 shows an example of a part of the input image. As shown in FIG. 5, white lines are formed in the partial image 201 in an oblique direction in the upper left and lower right. Such white lines form a similar image pattern in a plurality of blocks when viewed in a block narrower than the partial image 201. Thus, an image pattern in which the same pattern appears repeatedly is referred to as a repeated pattern.

  In the partial image 201, for example, a point on the white line is set as a target point 211, and a square area having a predetermined size centered on the target point 211 is set as a target block 212. For such a block of interest 212, the comparison target block 222 having the comparison point 221 as the center point and the comparison target block 232 having the comparison point 231 at the position different from the comparison point 221 have different images. Therefore, naturally, the degree of similarity with respect to the target block 212 is also different.

  The image evaluation processing unit 102 performs block matching between the comparison target block and the target block 212 using each pixel in the target block 212 as a comparison point, and indicates the similarity of the comparison target block to the target block 212 for each comparison point. An evaluation value is calculated. More specifically, in the target block 212 and the comparison target block, a difference between pixel values of pixels having the same in-block position is taken, and the sum of absolute values of the difference values is used as an evaluation value.

  A surface 240 of the graph shown in FIG. 6 shows the distribution of evaluation values calculated in this way. In FIG. 6, the size of the block of interest 212 is 31 pixels × 31 pixels. In this example, the smaller the evaluation value is, the higher the similarity is. For example, the point 241 indicates an evaluation value when the attention point 211 and the comparison point are the same, and in this case, the attention block 212 and the comparison target block are the same, so the evaluation value is “0”. ing.

  When the comparison point is located on the same white line as the attention point 211 in the image pattern such as the partial image 201 like the attention block 212 and the comparison target block 232 illustrated in FIG. 5, the comparison corresponding to the comparison point The composition of the image of the target block has a high similarity to the composition of the image of the block of interest 212. That is, since the evaluation value of such a comparison point becomes small, in the case of an image pattern such as the partial image 201, as shown in FIG. 6, a comparison with a low evaluation value in an oblique direction passing through the point of interest. The points will be distributed. In other words, such an evaluation value distribution pattern indicates the presence of a linear edge, that is, a repetitive pattern.

  Hereinafter, an example of such a method for detecting a linear edge will be described. FIG. 7 is a flowchart showing an example of the flow of straight edge detection processing executed by the repeated pattern detection unit 103. This linear edge detection process shows a specific example of the repeated pattern detection process described with reference to the flowchart of FIG. 4 (an example of detecting a linear edge as a repeated pattern).

  When the linear edge detection process is started, the feature point specifying unit 121 specifies a point having the highest similarity as a first feature point in a region separated by a predetermined distance or more from the target point in step S61.

  FIG. 8 is a diagram schematically showing the block of interest. The attention block 250 shown in FIG. 8 has a size of 15 pixels × 15 pixels for convenience of explanation. Note that the target block 250 may be considered to indicate a part near the center point of the target block 212 in FIG.

  As described above, the feature point specifying unit 121 first specifies the point having the smallest evaluation value (the point having the highest similarity) as the first feature point 253 in the block of interest 250. At this time, pixels in a predetermined range 252 centering on the target point 251 as the center point are excluded from the target of the first feature point. This is because the evaluation value in the vicinity of the point of interest 251 is generally small regardless of the presence or absence of the repeated pattern, and there is a risk of erroneous detection of the repeated pattern.

  When the first feature point 253 is specified, the feature point specifying unit 121 sets a first straight line 254 connecting the first feature point 253 specified by the process of step S61 and the point of interest 251 in step S62. .

  When the first straight line 254 is set, the feature point specifying unit 121 sets a second straight line 255 that passes through the point of interest 251 and is perpendicular to the first straight line 254 in step S63. In step S <b> 64, the feature point specifying unit 121 specifies two points on the second straight line 255 that are separated from the target point 251 by a predetermined distance in opposite directions as the second feature point 256 and the third feature point 257. Note that the second feature point 256 and the third feature point 257 are also specified in a region other than the range 252 (that is, near the point of interest 251) in the block of interest 250.

  The above steps S61 to S64 correspond to the process of step S41 in FIG.

  When the first feature point 253, the second feature point 256, and the third feature point 257 are specified, the feature analysis unit 122 determines that the second feature point 256 and the third feature point are in step S65. An average value of similarities corresponding to 257 (that is, evaluation values of the second feature point 256 and the third feature point 257) is calculated. This process corresponds to the process of step S42 in FIG.

  When the average value of the evaluation values of the second feature point 256 and the third feature point 257 is calculated, the analysis result determination unit 123 calculates the average value and the evaluation value of the first feature point 253 in step S66. A difference value is calculated, and it is determined whether or not the difference value is larger than a predetermined threshold value. That is, the analysis result determination unit 123 determines whether or not the similarity between the first feature points is higher than a predetermined threshold compared to the average value of the similarity between the second feature points and the third feature points. To do.

  If it is determined that the difference value is greater than the predetermined threshold value, the analysis result determination unit 123 proceeds to step S67, determines that a straight edge exists along the first straight line 254, and outputs the determination result. Then, the linear edge detection process is terminated, the process returns to step S3 in FIG. 2, and the pattern determination process is terminated. If it is determined that the difference value is not greater than the predetermined threshold value, the analysis result determination unit 123 proceeds to step S68, determines that there is no linear edge in the target block 250, and outputs the determination result. Then, the linear edge detection process is terminated, the process returns to step S3 in FIG. 2, and the pattern determination process is terminated.

  Note that the processing in steps S66 to S68 corresponds to the processing in step S43 in FIG.

  That is, when detecting the linear edge, the repeated pattern detection unit 103 connects the first feature point and the target point 251 that are determined to have a high similarity to the target block of the corresponding comparison target block (the first straight line). Assuming that a straight edge is present (along 254), two points (second feature point 256 and third feature point 257) in a direction (on the second straight line 255) perpendicular to the first straight line 254 are assumed. ) Are selected as two points that are not likely to be on the straight edge, and the average value of the evaluation values of these two points is compared with the evaluation value of the first feature point. At this time, the degree of similarity of the comparison block of the first feature point with respect to the target block is overwhelming with the degree of similarity of the comparison target block of each of the second feature point 256 and the third feature point 257 with respect to the target block. If it is high (that is, the difference between the average value of the evaluation values of the second feature point 256 and the third feature point 257 and the evaluation value of the first feature point is equal to or greater than the threshold value), the repeated pattern detection unit 103 determines that there is a straight edge along the first straight line 254.

  By performing the detection as described above, the repeated pattern detection unit 103 can easily and accurately detect a linear edge in an arbitrary direction included in the image.

  In the above, an example of the detection method of the linear edge has been described, but the linear edge may be detected by a method other than the above-described method.

  For example, four or more feature points may be specified, and a linear edge may be detected using a part or all of them, or a linear edge may be detected using two feature points. . Further, the second feature point 256 and the third feature point 257 have been described as two points on the second straight line 255 that are separated from each other by the predetermined distance from the point of interest 251. For example, The points not on the straight line 255 may be the second feature point 256 and the third feature point 257, and the distance between the second feature point 256 and the third feature point 257 to the point of interest 251 is mutually It may be different.

  Furthermore, the size and shape of the range 252 are arbitrary. In the above description, the evaluation value is described so as to be smaller as the similarity is higher. However, the evaluation value calculation method is arbitrary, and any parameter may be used as the evaluation value.

  In the above description, a case where a linear edge is detected is described as an example of detecting a repeated pattern. However, the repeated pattern detection unit 103 detects a geometric pattern such as a polka dot or a mesh as a repeated pattern. May be. The specific detection method in that case is arbitrary. Further, it is also possible to perform repeated pattern detection by a plurality of different methods and determine what pattern the detected repeated pattern is based on the detection results.

  When performing moving image pattern determination, the pattern determination apparatus 100 performs the above-described processes for each field or frame.

  By the way, in order to improve the accuracy of the repeated pattern detection, the repeated pattern detection described above is performed in each of a plurality of blocks of interest at different positions, and the repeated pattern is detected again based on the plurality of detection results. The presence or absence may be determined.

  FIG. 9 is a block diagram showing a configuration example of the pattern determination apparatus in that case. A pattern determination apparatus 300 shown in FIG. 9 is an apparatus that detects a repeated pattern in a plurality of blocks of interest. The pattern determination apparatus 300 has basically the same configuration as the pattern determination apparatus 100 in FIG. 1, but unlike the pattern determination apparatus 100, the image holding unit 101, the image evaluation processing unit 102, and the repeated pattern detection unit 103. In addition, a determination result holding unit 311, a new block setting unit 312, and a redetermination unit 313 are included.

  The determination result holding unit 311 has a storage medium such as a semiconductor memory, for example, and stores the determination result determined by the analysis result determination unit 123 in association with the target point or target block to be processed. Based on the determination result held in the determination result holding unit 311, the new block setting unit 312 specifies a new block for which the next repeated pattern detection is to be performed as a target block, performs the setting, and sets the image evaluation processing unit. The image evaluation (for example, block matching) of the identified new block is performed on the block 102.

  As described above, when the repeated pattern detection process is performed on a predetermined number of blocks of interest, the re-determination unit 313 is based on a part or all of the determination results held in the determination result holding unit 311. It is determined whether or not a repetitive pattern is included in the input image, and the determination result is output.

  Next, an example of the flow of the multiple block pattern determination process executed by the pattern determination apparatus 300 will be described with reference to the flowchart of FIG.

  When the multi-block pattern determination process is started, the image holding unit 101, the image evaluation processing unit 102, and the repeated pattern detection unit 103 first set the first block (first block) as a target block in step S101. Pattern determination processing is performed in the same manner as described with reference to the flowchart of FIG. When the repeated pattern detection unit 103 outputs the determination result, the determination result holding unit 311 holds the determination result in step S102. When the new block setting unit 312 determines in step S103 whether or not a repetitive pattern has been detected in the first block based on the determination result stored in the determination result storage unit 311 and determines that it has been detected. Then, the process proceeds to step S104.

  In step S <b> 104, the new block setting unit 312 controls the image evaluation processing unit 102 to set a block different from the block set as the target block so far as a new target block. In step S105, the image holding unit 101, the image evaluation processing unit 102, and the repetitive pattern detection unit 103 perform pattern determination processing on the new block of interest in the same manner as described with reference to the flowchart of FIG. . When the repeated pattern detection unit 103 outputs the determination result, the determination result holding unit 311 holds the determination result in step S106. In step S107, the new block setting unit 312 determines whether or not the pattern determination process has been repeated a predetermined number of times. If it is determined that the pattern determination process has not been repeated, the process returns to step S104 and the subsequent processes are repeated. If it is determined in step S107 that the pattern determination process has been repeated a predetermined number of times, the new block setting unit 312 advances the process to step S108. In other words, the pattern determination is performed for a predetermined number of blocks by repeatedly executing each processing from step S104 to step S107.

  In step S108, the re-determination unit 313 determines whether or not the input image includes a repetitive pattern based on the determination result of each block held in the determination result holding unit 311. For example, if all the repeated patterns detected in each block match and no contradiction occurs, the re-determination unit 313 determines that the repeated pattern is included in the input image. Note that this determination method is arbitrary, and any determination method may be used, and any determination criterion may be used.

  When the process of step S108 ends, the re-determination unit 313 ends the multiple block pattern determination process. If it is determined in step S103 that no repetitive pattern is detected in the first block, the new block setting unit 312 determines that the input image does not include a repetitive pattern, and ends the multiple block pattern determination process. .

  For example, as shown in FIG. 11, in the first block 410, it is determined that there is a straight line 413 (straight edge) passing through the point of interest 411 and the first feature point 412 of the first block 410. The new block setting unit 312 sets a point on the straight line 413 that is a predetermined distance (R2) away from the point of interest 411 as a new point of interest 421, and sets the second block as a new block of interest corresponding to the point of interest 421. 420 is set, and the second block 420 repeatedly performs pattern detection. Also, the new block setting unit 312 sets a point on the straight line 413 that is a predetermined distance (R3) away from the point of interest 411 as a new point of interest 431, and the third block as a new block of attention corresponding to the point of interest 431. 430 is set, and repeated pattern detection is performed in the third block of interest 430.

  Then, a straight line connecting the attention point 421 and the first feature point 422 detected in the second block 420 and a straight line connecting the attention point 431 and the first feature point 432 detected in the third block 430. However, if it matches or approximates the straight line 413, the re-determination unit 313 determines that the input image includes a linear edge (repeated pattern) along the straight line 413.

  As described above, by performing repeated pattern detection in a plurality of blocks, it is possible to improve the accuracy of determining a repeated pattern and more accurately detect a repeated pattern.

  Note that any block may be set as the target block for detecting this repeated pattern. For example, as shown in FIG. 12A, a plurality of blocks corresponding to the common attention point 441 and having different block sizes may be used as the attention block, such as the block 440, the block 450, and the block 460. Good. At this time, any of the block 440, the block 450, and the block 460 may be the first block.

  Also, for example, as shown in FIG. 12B, the pixels 471 in the first block are grouped into one pixel 472 by averaging the pixel values for each predetermined number of pixels. A reduced image is created, the reduced image is used as a second block, and each pixel value of the pixels 472 in the second block is averaged for each predetermined number of pixels. In summary, a reduced image of the second block may be created, the reduced image may be used as the third block, and a repeated pattern may be detected for each.

  Of course, other methods may be used. For example, a suitable method for selecting a target block differs depending on the repeated pattern to be detected. Therefore, this selection method may be selected according to the repeated pattern to be detected.

  In the above description, it has been described that when a repetitive pattern is detected in the first block, the repetitive pattern is detected in a plurality of blocks. However, the present invention is not limited to this, and the processing in step S103 is omitted. Regardless of the result, it is also possible to detect a repeated pattern in a plurality of blocks. Further, the re-determination unit 313 may adopt the most likely detection result even if the detection results in the plurality of blocks do not all match or approximate. By doing in this way, the pattern determination apparatus 300 can improve the robustness of repeated pattern detection.

  Note that the number of blocks in which the repeated pattern is detected is arbitrary.

  Alternatively, a plurality of blocks may be set as the first block, and the above-described multiple block pattern determination process may be repeatedly executed. At this time, by making different blocks to be processed in each of the multiple block pattern determination processes, it is possible to increase the number of blocks to be subjected to repeated pattern detection, and to improve the repeat pattern determination accuracy and repeat pattern detection robustness. Further improvement can be made and the repeated pattern can be detected more accurately.

  Further, when comparing the blocks, the block matching unit 115 may divert the past comparison result without actually comparing the blocks if the comparison between the same blocks has been performed in the past. Good.

  For example, in the example of FIG. 5, the evaluation value of the comparison point 231 is the same as the evaluation value of the comparison point 211 when the comparison point 231 is the attention point and the attention point 211 is the comparison point (similarity does not change). ). Therefore, for example, the evaluation value holding unit 117 associates the evaluation value with a combination of the attention point and the comparison point (a combination of coordinate information of two points, which is the attention point and which is the comparison point), First, the block matching unit 115 searches the evaluation value of the evaluation value holding unit 117 from the combination of the attention point and the comparison point to be compared, and when there is an evaluation value corresponding to the combination, the evaluation value is adopted, Block matching by the block matching unit 115 and calculation of the evaluation value by the evaluation value calculation unit 116 may be omitted.

  By doing so, the processing load can be reduced.

  Further, for example, when the block matching unit 115 calculates a difference value between pixels in the block matching process, all the difference values are held, and in the block matching process executed thereafter, the same pixel value When the difference value is calculated, that value may be used. By doing so, the processing load can be reduced.

  Note that, when performing moving image pattern determination, the pattern determination apparatus 300 performs the above-described processes for each field or frame.

  By the way, the pattern determination apparatus 100 or the pattern determination apparatus 300 as described above may be used as a pattern determination unit, and may be incorporated in an apparatus that performs other processing.

  FIG. 13 is a block diagram illustrating a configuration example of a monitoring camera device to which the present invention is applied. A monitoring camera device 500 shown in FIG. 13 is a device used to monitor a predetermined room, area, or the like with a video, and uses the above-described repeated pattern detection result for a captured image for camera control processing. Device.

  The monitoring camera device 500 causes the image display 512 to display an image captured by the image capturing unit 511 including, for example, a CCD video camera. The user (monitoring person) grasps (monitors) the state of the space captured by the imaging unit 511 by browsing the image displayed on the image display 512.

  The tracking processing unit 513 detects a predetermined tracking target in the image input by the imaging unit 511 and supplies the detection result to the camera driving unit 514. The camera driving unit 514 drives the imaging unit 511 so as to follow the tracking target, and performs, for example, zoom and focus adjustment, tilt and pan, and still image shooting.

  The tracking processing unit 513 includes a pattern determination unit 521 and a block matching processing unit 522. The pattern determination unit 521 has the same configuration as the pattern determination device 100 or the pattern determination device 300 described above, performs the same processing, and generates a repetitive pattern such as a linear edge from the image input by the imaging unit 511. To detect.

  The block matching processing unit 522 tracks the tracking target by performing block matching between the frame images continuously input by the imaging unit 511 (the position of the tracking target in the new frame image is the previous one). Specified by block matching with the frame image). At this time, for example, when a repetitive pattern exists in the frame image, the tracking target is located near the repetitive pattern, and the block including the tracking target of the comparison source includes the repetitive pattern, the plurality of blocks including the repetitive pattern include Therefore, there is a possibility that accurate matching cannot be performed.

  Therefore, the block matching processing unit 522 performs block matching with reference to the detection result of the repeated pattern by the pattern determining unit 521, and if the repeated pattern is included in the block including the comparison target tracking target, the repeated processing is performed. In order not to be affected by the pattern, the position of the tracking target is specified by a method different from the normal case. By doing so, the tracking processing unit 513 can track the tracking target more accurately, and the camera driving unit 514 can drive the imaging unit 511 more appropriately. That is, the monitoring camera device 500 can perform more appropriate monitoring behavior.

  Of course, the pattern determination processing unit to which the present invention is applied can be applied to any device other than the monitoring camera device as long as it performs processing using the repeated pattern detection result.

  The series of processes described above can be executed by hardware or can be executed by software. In this case, for example, it may be configured as a personal computer as shown in FIG.

  In FIG. 14, a CPU (Central Processing Unit) 601 of the personal computer 600 performs various processes according to a program stored in a ROM (Read Only Memory) 602 or a program loaded from a storage unit 613 into a RAM (Random Access Memory) 603. Execute the process. The RAM 603 also appropriately stores data necessary for the CPU 601 to execute various processes.

  The CPU 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input / output interface 610 is also connected to the bus 604.

  The input / output interface 610 includes an input unit 611 including a keyboard and a mouse, a display including a CRT (Cathode Ray Tube) and an LCD (Liquid Crystal Display), an output unit 612 including a speaker, a hard disk, and the like. A communication unit 614 including a storage unit 613 and a modem is connected. The communication unit 614 performs communication processing via a network including the Internet.

  A drive 615 is connected to the input / output interface 610 as necessary, and a removable medium 621 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory is appropriately mounted, and a computer program read from them is loaded. Installed in the storage unit 613 as necessary.

  When the above-described series of processing is executed by software, a program constituting the software is installed from a network or a recording medium.

  For example, as shown in FIG. 14, the recording medium is distributed to distribute the program to the user separately from the apparatus main body, and includes a magnetic disk (including a flexible disk) on which the program is recorded, an optical disk ( It only consists of removable media 621 consisting of CD-ROM (compact disc-read only memory), DVD (including digital versatile disc), magneto-optical disc (including MD (mini disc)), or semiconductor memory. Rather, it is composed of a ROM 602 storing a program and a hard disk included in the storage unit 613, which is distributed to the user in a state of being incorporated in the apparatus main body in advance.

  In the present specification, the step of describing the program recorded on the recording medium is not limited to the processing performed in chronological order according to the described order, but is not necessarily performed in chronological order. It also includes processes that are executed individually.

  Further, in this specification, the system represents the entire apparatus constituted by a plurality of devices (apparatuses).

  In the above, the configuration described as one device may be divided and configured as a plurality of devices. Conversely, the configurations described above as a plurality of devices may be combined into a single device. Of course, configurations other than those described above may be added to the configuration of each device. Furthermore, if the configuration and operation of the entire system are substantially the same, a part of the configuration of a certain device may be included in the configuration of another device. That is, the embodiment of the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the gist of the present invention.

It is a block diagram which shows the main structural examples of the pattern determination apparatus to which this invention is applied. It is a flowchart explaining the example of the flow of a pattern determination process. It is a flowchart explaining the example of the flow of an image evaluation process. It is a flowchart explaining the example of the flow of a repeating pattern detection process. It is a figure explaining the example of the mode of block matching. It is a figure which shows the example of the mode of distribution of evaluation value. It is a flowchart explaining the example of the flow of a linear edge detection process. It is a figure explaining the example of the mode of a linear edge detection. It is a block diagram which shows the other structural example of the pattern determination apparatus to which this invention is applied. It is a flowchart explaining the example of the flow of a multiple block pattern determination process. It is a figure explaining the example of the mode of the pattern determination process using several blocks. It is a figure explaining the example of the selection method of multiple blocks. It is a block diagram which shows the main structural examples of the surveillance camera apparatus to which this invention is applied. It is a block diagram which shows the structural example of the personal computer to which this invention is applied.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 pattern determination apparatus, 101 image holding part, 102 image evaluation process part, 103 repetitive pattern detection part, 111 attention point specification part, 112 attention block specification part, 113 comparison point selection part, 114 comparison object block specification part, 115 block matching Unit, 116 evaluation value calculation unit, 121 feature point specification unit, 122 feature analysis unit, 123 analysis result determination unit, 300 pattern determination device, 311 determination result holding unit, 312 new block setting unit, 313 re-determination unit

Claims (11)

An information processing apparatus for detecting a repetitive pattern in which a predetermined image pattern included in an image is repeated,
Image evaluation means for evaluating each pixel in a target block that is a predetermined area corresponding to a target point that is a target pixel of the image, and obtaining each evaluation value;
In the target block, feature point specifying means for specifying a plurality of feature points that are pixels other than the target point for analyzing the distribution of the evaluation value in the target block;
An information processing apparatus comprising: a determination unit that determines whether or not the image includes the repetitive pattern based on a relationship between evaluation values of the plurality of identified feature points. The evaluation value is obtained by comparing the target block with a target block that is a predetermined area corresponding to a comparison point that is a pixel in the target block selected as a target for comparison. The information processing device according to claim 1, wherein the information processing device is a parameter indicating a similarity to the block of interest. The evaluation value of the comparison point is a sum of absolute values of difference values between pixels of the comparison target block corresponding to the comparison point and the block of interest in the same position in the block. Information processing device. The information processing apparatus according to claim 1, wherein the feature point specifying unit specifies three points as the feature points. The repeating pattern is an image of a straight edge,
The information processing apparatus according to claim 1, wherein the determination unit determines whether or not there is an image of the linear edge, and also determines the direction of the linear edge when determining that the image of the linear edge exists. . The feature point specifying means specifies, as a first feature point, a comparison point of an evaluation value indicating that the similarity of the corresponding comparison target block to the target block is the highest except for the target point, and 2. Two points on a straight line orthogonal to a straight line passing through the first feature point and the point of interest, which are separated from the point of interest by a predetermined distance in a direction opposite to each other, are specified as a second feature point and a third feature point. Item 6. The information processing device according to Item 5. The feature point specifying means specifies the first feature point to the third feature point from pixels in a region excluding a range within a predetermined distance from the target point in the target block. The information processing apparatus described. A feature analysis means for analyzing a feature of the distribution of the evaluation values by calculating an average value of the evaluation values of the first feature points and the second feature points;
The determination unit compares the average value calculated by the feature analysis unit with the evaluation value of the first feature point, and if the difference is larger than a predetermined threshold, the first feature point and the attention point The information processing apparatus according to claim 7, wherein it is determined that the linear edge exists along a straight line passing through a point. Re-determination means for determining again whether or not the image includes the repetitive pattern based on the determination result by the determination means;
The image evaluation means obtains the evaluation value for each pixel of a plurality of blocks of interest,
The feature point specifying means specifies a plurality of feature points in each of the plurality of blocks of interest,
The determination means determines, for each of the plurality of blocks of interest, whether the image includes the repeating pattern,
The information processing apparatus according to claim 1, wherein the re-determination unit determines again whether or not the image includes the repetitive pattern based on a plurality of determination results obtained by the determination by the determination unit. An information processing method of an information processing apparatus for detecting a repetition pattern in which a predetermined image pattern included in an image is repeated,
Evaluate each pixel in the block of interest that is a predetermined area corresponding to the point of interest that is the pixel to be processed of the image, to obtain each evaluation value,
In the target block, a plurality of feature points that are pixels other than the target point for analyzing the distribution of the evaluation value in the target block are specified,
An information processing method including a step of determining whether or not the image includes the repetitive pattern based on a relationship between evaluation values of a plurality of identified feature points. In a program for detecting a repetitive pattern in which a predetermined image pattern included in an image is repeated,
An image evaluation step for evaluating each pixel in a target block that is a predetermined area corresponding to a target point that is a pixel to be processed of the image, and obtaining each evaluation value;
In the target block, a feature point specifying step for specifying a plurality of feature points that are pixels other than the target point for analyzing the distribution of the evaluation value in the target block;
A program for causing a computer to execute a determination step of determining whether or not the image includes the repetitive pattern based on a relationship between evaluation values of a plurality of identified feature points.

Images