JP4882929B2 - Image processing apparatus and image processing program - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing program.

  When a handwritten character is to be recognized, the recognition rate decreases if there is a contact with a plurality of characters or other figures, so that it is required to separate them.

  As a technique related to this, for example, Patent Document 1 aims at easily recognizing a character string regardless of the number of contact characters and the way of contact using structural information of the entire character string image. In the column recognition process, first, the extracted character string image is thinned to generate a skeleton line figure, then the skeleton line figure is divided into a plurality of simple partial curves, and then the partial curves are combined to create a character. Generate a partial figure of the column, perform character recognition on the partial figure, then make the partial figure recognized by the character a node, connect the node with weighted links variously to form a network, and finally , Selecting an optimal solution route from a network and outputting a combination of the nodes.

  For example, in Patent Document 2, it is an object to improve processing efficiency by suppressing cutting candidates and the number of cuttings when a stroke extracted from a character image is a contact stroke, and an initial image is input from a character string image storage unit. The stroke extraction unit extracts the initial stroke, the cutting candidate extraction unit extracts the cutting candidate location only for the stroke determined as the contact stroke determination unit, the optimum candidate selection unit selects the cutting location from the cutting candidate, and the stroke The information update unit cuts the contact stroke at the selected location to generate a new stroke, and the character cutout candidate generation unit combines the new stroke and the non-contact stroke stored in the non-contact stroke buffer to generate a character. It is disclosed that a cutout candidate is generated and output.

In addition, for example, Patent Document 3 can generate a character extraction candidate that includes a correct answer with high probability even if there is a contact portion where the stroke completely intersects without being affected by noise and line width fluctuation, In addition, an object is to provide a character segmentation candidate generation device that can prevent an increase in the number of candidates. A character string image storage unit stores an optically scanned character string image, and a line segment extraction unit stores this character string image. A line segment, which is a part having a specific direction component, is extracted from the extracted character string image using a morphological operation, and the stroke creation unit merges the extracted line segments and connects them to some extent smoothly. It is disclosed that a stroke that is a part is created, and finally, a character segmentation candidate creation unit creates a character segmentation candidate by combining the created strokes.
Japanese Patent Laid-Open No. 06-076114 JP-A-11-149521 JP 09-288715 A

By the way, when the line figure is extracted by limiting the height or the contact position of the character image, it cannot be dealt with if there is a size, contact or intersection of other line figures.
The present invention generates a graph model, and extracts an open figure and a closed figure separately, thereby relaxing the variation in the size of the line figure, the restriction of contact or intersection, and extracting the line figure. An object is to provide an apparatus and an image processing program.

The gist of the present invention for achieving the object lies in the inventions of the following items.
First aspect of the present invention, and an extraction means for extracting features and line segment of the point in the image containing the line figure, the thus extracted characteristic points to the extraction means and the vertices in the graph model, extracted by said extracting means By making the obtained line segment an edge in the graph model, a graph model generating unit that generates a graph model, and an open figure having an end point are extracted by analyzing the graph model generated by the graph model generating unit and an open figure extraction means, by analyzing the graph model generated by the graph model generating means, comprising a closed figure extracting means for extracting closed figure having no end points, the closed figure extracting means extracts the closed An image processing apparatus is characterized in that an overlapping graphic is deleted from the graphic .
The invention according to claim 2 is an extraction means for extracting features and line segments related to points in an image including a line figure, and the feature points extracted by the extraction means are vertices in a graph model, and is extracted by the extraction means. By defining the line segment as an edge in the graph model, the graph model generating means for generating the graph model and the open model having the end points are extracted by analyzing the graph model generated by the graph model generating means. A figure extraction means; a closed figure extraction means for extracting a closed figure having no end points by analyzing the graph model generated by the graph model generation means; an open figure extracted by the open figure extraction means; A predetermined figure in which the total of the closed figures extracted by the closed figure extraction means is included in the image. A determination means for determining whether or not the number is different, and if determined to be different by the determination means, the graph model is changed according to the location where the line segment is in contact in the image, An image processing apparatus comprising reprocessing means for performing processing by an open figure extraction means or the closed figure extraction means.
According to a third aspect of the present invention, an extraction means for extracting features and line segments related to points in an image including a line figure, and a feature point extracted by the extraction means is used as a vertex in a graph model, and is extracted by the extraction means. By defining the line segment as an edge in the graph model, the graph model generating means for generating the graph model and the open model having the end points are extracted by analyzing the graph model generated by the graph model generating means. A figure extraction means, a closed figure extraction means for extracting a closed figure having no end points by analyzing the graph model generated by the graph model generation means, and an open figure extracted by the open figure extraction means or the A predetermined figure in which the shape of the closed graphic extracted by the closed graphic extracting means is included in the image. If it is determined by the determination means that determines whether the shape is different from the shape, the graph model is changed according to the location where the line segment is in contact in the image, An image processing apparatus comprising reprocessing means for performing processing by an open figure extraction means or the closed figure extraction means.

The invention according to claim 4 comprises false line segment identifying means for identifying a false line segment out of the line segments based on the feature points and line segments extracted by the extracting means, and the graph model generating means is based on the result of the extraction means and the false line identification means is an image processing apparatus according to any one of claims 1 to 3, and generates the graph model.

The invention according to claim 5 is characterized in that the closed graphic extracting means recognizes the shape of the closed graphic after extracting the closed graphic and performing polygon approximation by analyzing the graph model. The image processing apparatus according to claim 1.

According to a sixth aspect of the invention, a computer, extracting means for extracting a feature and line segment of the point in the image containing the line figure, the vertices in the graph model thus extracted feature points to the extraction means, the extraction A graph model generating means for generating a graph model by using the line segment extracted by the means as an edge in the graph model, and an open figure having end points by analyzing the graph model generated by the graph model generating means By analyzing the graph model generated by the graph model generating means, and functioning as a closed figure extracting means for extracting a closed figure having no end points . An image processing program that deletes an overlapping figure from among the extracted closed figures .
According to the seventh aspect of the present invention, there is provided an extraction means for extracting a feature and a line segment related to a point in an image including a line figure, a feature point extracted by the extraction means as a vertex in a graph model, and the extraction means By using the line segment extracted by the graph as an edge in the graph model, a graph model generation unit that generates a graph model, and an open figure having end points by analyzing the graph model generated by the graph model generation unit An open graphic extraction means for extracting; a closed graphic extraction means for extracting a closed graphic having no end points by analyzing the graph model generated by the graph model generation means; and an open graphic extracted by the open graphic extraction means. The total of the figure and the closed figure extracted by the closed figure extraction means is included in the image. A determination means for determining whether or not the number of figures is different from the number of figures determined, and the graph model when determined to be different by the determination means, depending on a position where the line segment is in contact in the image The image processing program is made to function as a reprocessing unit that performs processing by the open graphic extracting unit or the closed graphic extracting unit.
According to an eighth aspect of the present invention, there is provided an extraction means for extracting a feature and a line segment related to a point in an image including a line figure, and a feature point extracted by the extraction means as a vertex in a graph model. By using the line segment extracted by the graph as an edge in the graph model, a graph model generation unit that generates a graph model, and an open figure having end points by analyzing the graph model generated by the graph model generation unit An open graphic extraction means for extracting; a closed graphic extraction means for extracting a closed graphic having no end points by analyzing the graph model generated by the graph model generation means; and an open graphic extracted by the open graphic extraction means. The figure or the shape of the closed figure extracted by the closed figure extraction means is included in the image. A determination means for determining whether or not the shape of the figure is different from the shape of the figure determined, and if it is determined by the determination means, the graph model according to a position where the line segment is in contact in the image The image processing program is made to function as a reprocessing unit that performs processing by the open graphic extracting unit or the closed graphic extracting unit.

According to the image processing apparatus of the first aspect, it is possible to extract the line figure by relaxing the variation in the size of the line figure, the restriction on the contact or the intersection. Furthermore, it becomes possible to suppress extraction of extra line figures.
According to the image processing apparatus of the second aspect, it is possible to extract the line figure while relaxing the variation in the size of the line figure, the restriction on the contact or the intersection. Furthermore, when the number of figures is known in advance, a line figure can be extracted more accurately.
According to the image processing apparatus of the third aspect, it is possible to extract the line figure while relaxing the variation in the size of the line figure, the restriction on the contact or the intersection. Further, when the shape of the figure is known in advance, the line figure can be extracted more accurately.

According to the image processing apparatus of the fourth aspect , it is possible to identify a false line segment that is likely to be generated at the intersecting portion based on the thickness of the line segment, and it is possible to extract a more accurate line figure.

According to the image processing apparatus of the fifth aspect , it becomes possible to extract a more accurate line figure by recognizing the shape.

According to the image processing program of the sixth aspect , it is possible to extract the line figure by relaxing the variation in the size of the line figure, the restriction on the contact or the intersection. Furthermore, it becomes possible to suppress extraction of extra line figures.
According to the image processing program of the seventh aspect, it is possible to extract the line figure by relaxing the variation in the size of the line figure, the restriction on the contact or the intersection. Furthermore, when the number of figures is known in advance, a line figure can be extracted more accurately.
According to the image processing program of the eighth aspect, it is possible to extract the line figure by relaxing the variation in the size of the line figure, the restriction on the contact or the intersection. Further, when the shape of the figure is known in advance, the line figure can be extracted more accurately.

Hereinafter, an example of a preferred embodiment for realizing the present invention will be described with reference to the drawings.
FIG. 1 shows a conceptual module configuration diagram of a configuration example of the present embodiment.
The module generally refers to components such as software (computer program) and hardware that can be logically separated. Therefore, the module in the present embodiment indicates not only a module in a computer program but also a module in a hardware configuration. Therefore, the present embodiment also serves as an explanation of a computer program, a system, and a method. However, for the sake of explanation, the words “store”, “store”, and equivalents thereof are used, but these words are stored in a storage device when the embodiment is a computer program, or It means to control to be stored in the storage device. In addition, the modules correspond almost one-to-one with the functions. However, in mounting, one module may be composed of one program, or a plurality of modules may be composed of one program. A plurality of programs may be used. The plurality of modules may be executed by one computer, or one module may be executed by a plurality of computers in a distributed or parallel environment. Note that one module may include other modules. In the following, “connection” includes not only physical connection but also logical connection (data exchange, instruction, reference relationship between data, etc.).
In addition, the system or device is configured by connecting a plurality of computers, hardware, devices, and the like by communication means such as a network (including one-to-one correspondence communication connection), etc., and one computer, hardware, device. The case where it implement | achieves by etc. is also included.

As shown in FIG. 1, the present embodiment includes a preprocessing module 11, a line tracking extraction module 12, a shape analysis extraction module 13, and a postprocessing module 14.
The preprocessing module 11 is connected to the line tracking extraction module 12, inputs a target image, performs preprocessing on the image, and passes the image to the line tracking extraction module 12.
Here, the target image is an image including a plurality of line figures entered by handwriting. The line figure may be an image (position data as a trajectory) including a line figure electronically written with an electronic pen or the like in addition to a paper document handwritten with a pen. The paper document image is input by an image input device such as a scanner. More specifically, for example, an image such as a handwritten line figure added to an answer sheet on which a test for students / students has been determined, a questionnaire sheet on which an answer has been written, and the like is applicable.
Preprocessing includes noise removal processing and thinning processing. Examples of the noise removal processing include removal of isolated points, interpolation of blurred lines, and inclination correction. Thinning processing refers to processing that converts a binary image into a line image with a width of one pixel. Processing used as preprocessing such as pattern recognition (matching with a predetermined pattern is performed, and if they match, Etc.) are used.

The line tracking extraction module 12 is connected to the preprocessing module 11 and the shape analysis extraction module 13, receives the line figure image that has been thinned from the preprocessing module 11, and tracks and extracts the line segments therein. By doing so, an open figure (figure with end points) is extracted. The processing of the line tracking extraction module 12 will be described later with reference to FIG.
The shape analysis extraction module 13 is connected to the line tracking extraction module 12 and the post-processing module 14. The shape analysis extraction module 13 receives an image from which an open figure has been extracted by the line tracking extraction module 12, and a closed figure (endpoint is an end line figure). No shape). The analysis result is passed to the post-processing module 14. The processing of the shape analysis extraction module 13 will be described later with reference to FIG.

The post-processing module 14 is connected to the shape analysis extraction module 13, receives the analysis result from the shape analysis extraction module 13, compares the total number of open figures and closed figures with the number of line figures known in advance, Or, if the shape of the line figure that is the analysis result is compared with the shape of the line figure that is known in advance, it is judged that they are different. 12 or the graph model used in the shape analysis extraction module 13 is changed, and control is performed so that the line tracking extraction module 12 or the shape analysis extraction module 13 performs the process again. Note that the number or shape of line figures known in advance is, for example, the answer sheet, questionnaire sheet, etc. The number and shape of handwritten symbols to be added may be determined in advance. It is. The processing of the post-processing module 14 will be described later with reference to FIG. Note that the processing by the post-processing module 14 is not necessarily performed.
That is, in this embodiment, the post-processing module 14 outputs individual line figures (images or shapes, positions, numbers, etc.) for an image including a plurality of line figures received by the pre-processing module 11. Will do.

FIG. 2 is an explanatory diagram showing the entire processing including the present embodiment.
This embodiment is used for, for example, scoring of answer sheets that have been tested for students, students, etc., and counting questionnaire sheets on which answers have been described. Therefore, it is necessary to extract a line figure added by handwriting.
As the reception image 21 which is a target image, for example, an answer sheet that has been tested for students, students, etc., a questionnaire sheet in which responses have been entered, and the like are applicable. In the received image 21, for example, as shown in FIG. 2, handwritten line figures of circles, triangles, and check symbols are drawn as overlapping line figures, and they overlap each other.
The postscript extraction module 22 extracts a line figure added from the received image 21 and generates a postscript image 23. For example, the postscript image 23 is generated by taking the difference from the original image of the received image 21 (that is, an image that has not been added).
The one-stroke line figure separation module 24 includes the configuration shown in FIG. 1, and separates the one-stroke line image 25, 26, 27 that is a one-stroke line figure from the postscript image 23. Here, the line drawing drawn by one stroke is not limited to the line drawing drawn by one stroke, but includes a line figure similar to this (for example, a simple figure such as “×”). Conversely, even if two line figures drawn with a single stroke (for example, a line figure connecting a circle and a triangle) are separated as a line figure (in the previous example, a circle and a triangle) To separate.

FIG. 3 is a conceptual module configuration diagram of a configuration example of the preprocessing module 11 and the line tracking extraction module 12.
The preprocessing module 31 corresponds to the preprocessing module 11, and the line tracking extraction module 12 is a feature point / line segment extraction module 32, a false line segment extraction module 33, a graph model construction module 34, a node analysis module 35, and an open figure extraction module. 36.
The preprocessing module 31 is connected to the feature point / line segment extraction module 32, and is equivalent to the preprocessing module 11 described with reference to FIG.

The feature point / line segment extraction module 32 is connected to the preprocessing module 31 and the false line segment extraction module 33, and is included in an image (an image including a plurality of line figures) preprocessed by the preprocessing module 31. Extract feature points and line segments for the points.
For example, a 3 × 3 pixel or a 5 × 5 pixel is extracted as a feature point related to a point, and is compared with a pattern of 3 × 3 pixels or 5 × 5 pixels previously determined as end points, intersections, or branch points. When a matching pattern is detected, the pixel is set as an end point, an intersection, or a branch point.
Then, the remaining pixels are traced and connected to form a line segment.

  The false line segment extraction module 33 is connected to the feature point / line segment extraction module 32 and the graph model construction module 34, and based on the feature points and line segments extracted by the feature point / line segment extraction module 32, a line is extracted. Extract false line segments from minutes.

An example of processing performed by the false line segment extraction module 33 will be described with reference to FIG.
FIG. 5A is an image received by the preprocessing module 31 and is an image in which a triangle and a quadrangle intersect.
FIG. 5B is an image that has been thinned by the preprocessing module 31. The portion of the target line segment 51 shown in FIG. 5B is a false line segment. Here, the false line segment refers to a line segment generated at a contact or intersection of line figures, and a portion that originally intersects at a point is a portion that is a line segment by thinning processing.
FIG. 5C is an enlarged view of the portion of the target line segment 51 of FIG. 5B superimposed on the corresponding portion of FIG. That is, a line segment having a width represented by crossing diagonal lines in FIG. 5C (portion surrounded by the outline 52 of the line segment) is the line segment in FIG. The central line segment (thinned line segment 53, attention line segment 54, etc.) in FIG. 5C is the thinned line segment in FIG.

An example of processing by the false line segment extraction module 33 is as follows.
(1) The average of the line widths 55 before the thinning process is obtained.
(2) Extract a short line segment with respect to the average line width. In the example of FIG. 5C, the attention line segment 54 is extracted. This is because the false line segment is generated at a place that should originally be an intersection or the like, and therefore, the fact that the line segment length is short is used. Judgment whether or not the line segment is a short line segment is based on comparison with a value obtained by multiplying the obtained line width average by a constant.
(3) When the shortest distance between each pixel in the line segment of interest 54 and the outline 52 of the line segment is large, the line segment is extracted as a false line segment. This is based on the fact that the false line segment tends to be a place where the line segment is thick. Whether or not the shortest distance is large is determined by comparison with a threshold (predetermined value, 1/2 of average line width, etc.).
(4) Further, when both ends of the line segment are odd-numbered branches and are not end points, the conditions (comparison with a threshold value, etc.) are made stricter (easily determined as a false line segment), and (2), (3 ) May be performed. This utilizes the fact that when the intersection is a false line segment, both ends of the false line segment become branch points.
(5) Then, the center of the extracted false line segment is changed as the feature point extracted by the feature point / line segment extraction module 32. Also, the thinned figure is corrected so that the line segments intersect at the feature points.

The graph model construction module 34 is connected to the false line segment extraction module 33 and the node analysis module 35, and generates a graph model based on the results of the feature point / line segment extraction module 32 and the false line segment extraction module 33. .
Note that the graph model construction module 34 may generate a graph model based on the result of the feature point / line segment extraction module 32 without performing the processing by the false line segment extraction module 33.
The processing of the graph model construction module 34 will be described later with reference to FIGS.

  The node analysis module 35 is connected to the graph model construction module 34 and the open figure extraction module 36, and analyzes a node of each figure based on the graph model generated by the graph model construction module 34. Specifically, step S412 to step S422 of the flowchart shown in FIG. 4 will be described later.

The open figure extraction module 36 is connected to the node analysis module 35 and extracts line figures that are not closed.
The open figure extraction module 36 may extract the open figure by analyzing the graph model generated by the graph model construction module 34 without performing the processing by the node analysis module 35.

An example of processing by the open figure extraction module 36 is as follows.
(1) An odd node (a node having an odd number of sides (edges) connected to the node) is set as a starting point. That is, the end point or the branch point is set as the start point.
(2) Tracking is performed on the thinned image. When there are multiple branches, the line segment with the smallest change in angle (that is, the angle formed by the line segments is close to 180 degrees) is tracked.
(3) When the pixel that has already been tracked is reached, the process ends. The trace traced from the start point to the end point is an open figure, and the open figure is extracted thereby.

FIG. 4 is a flowchart showing an example of processing performed by the preprocessing module 11 and the line tracking extraction module 12.
In step S402, the preprocessing module 31 receives an image including a line figure.
In step S404, the preprocessing module 31 performs preprocessing (noise removal processing and thinning processing) as described above.
In step S406, the feature point / line segment extraction module 32 performs the feature point / line segment extraction process as described above.
In step S408, the false line segment extraction module 33 extracts false line segments as described above.
In step S410, the graph model construction module 34 generates a graph model using the results obtained in steps S406 and S408. This process will be described later with reference to FIGS.

The processing from step S412 to step S422 is processing performed by the node analysis module 35.
In step S412, the number of nodes adjacent to each node (connected by edges) in the graph model generated in step S410 is calculated.
In step S414, it is determined whether the number of adjacent nodes of the target node is 3 or more. That is, it is determined whether it is an end point or a bending point. When it is 3 or more (when it is a branch point or an intersection), the process proceeds to step S418, and when it is less than 3 (when it is an end point or a bending point), the process proceeds to step S416.
In step S416, it is determined whether all nodes have been analyzed. If all the nodes have been analyzed, the process proceeds to S424. Otherwise, the process returns to step S414 for the next node.

In step S418, it is determined whether the number of nodes is an even number. If it is an even number, the process proceeds to step S420, and if it is an odd number, the process proceeds to step S422.
In step S420, even-numbered nodes are analyzed. In other words, you will find a pair of sides. Next, the process proceeds to step S416.
In step S422, an odd node is analyzed. In other words, in addition to the paired sides, the sides that are not paired are discovered. Next, the process proceeds to step S416.
Steps S424 and S426 are performed by the open graphic extraction module 36.
In step S424, it is determined whether there is an odd node that is not tracked. If it does not exist, the process ends (step S428), and if it exists, the process proceeds to step S426.
In step S426, the open figure extraction module 36 extracts an open figure as described above.

The process of the graph model construction module 34 will be described with reference to FIGS.
FIG. 6A shows a thinned line figure 600 as a result of thinning processing performed by the preprocessing module 31. This thin line figure 600 is a handwritten character “4”. Here, as a coordinate system, the upper left of the image is defined as the origin (0, 0), the X axis on the right and the Y axis on the lower side.
FIG. 6B shows feature points (a, b, c, d, e in the figure) and line segments extracted by the feature point / line segment extraction module 32 and the false line segment extraction module 33. Is also called a vertex. A line segment connecting these nodes is also called a side (edge).

FIG. 7 is a flowchart illustrating an example of processing performed by the graph model construction module 34 and the node analysis module 35. Note that what is surrounded by a dotted line on the right side of the flowchart indicates data to be output (or input) at each step. For example, the vertex is data output in step S702 and used in step S704.
Here, a vertex information table 80 that stores data about the vertices (feature points) of the line graphic and a side information table 90 that stores data about the edges of the line graphic will be described. The flowchart shown in FIG. 7 is a process for filling cells in the vertex information table 80 and the edge information table 90. FIG. 8 is an explanatory diagram showing the data structure of the vertex information table 80. The vertex information table 80 has a vertex column 81, an X coordinate column 82, a Y coordinate column 83, and a connection destination vertex column 84. FIG. 9 is an explanatory diagram showing the data structure of the side information table 90. The side information table 90 includes a side column 91, an ac column 92 to an ec column 95, and a side column 96 that is a selected set, and is used for analyzing the side connected to the vertex c. However, the number in the ac column 92 to the ec column 95 is the number of sides connected to the vertex c.

  In step S702, vertices are extracted. The coordinates of the feature points are stored in the X coordinate column 82 and the Y coordinate column 83 of the vertex information table 80 based on the result of the previous module. For example, since the X coordinate of the vertex a is 10 and the Y coordinate is 100, 10 and 100 are stored in the X coordinate column 82 and the Y coordinate column 83 corresponding to the vertex a, respectively.

  In step S704, the connection destination is searched. In other words, a line segment is traced from each vertex and another vertex is searched. Here, using the vertices extracted in step S702, the line segment connected to the vertices is traced to search for a connected vertex, and the result is stored in the connected vertex column 84 of the vertex information table 80. . For example, when the line figure in FIG. 6B is traced as the connection destination of the vertex a, the vertex b and c are reached, and the vertex b and c are stored in the connection destination vertex column 84 corresponding to the vertex a.

  In step S706, a vertex side having three or more connection destinations is extracted. That is, since the connection destination of each vertex is found in step S704, the side is determined using this. For example, since there are vertices a, b, c, and d as the connection destination of the vertex c, the line segments of ac, bc, dc, and ec are determined as sides. For example, the ac column 92 to the ec column 95 (that is, the number of columns) of the side information table 90 are determined.

  In step S708, an angle formed by the sides at a vertex having three or more sides is obtained. For example, since the angle formed by the side ac and the side bc is 100 degrees, 100 is stored in the ac column 92 and the bc column 93 corresponding to the sides bc and ac in the side information table 90.

  In step S710, a side to be paired is determined using the angle obtained in step S708. Here, the pair of sides is a side connected to the target side, and an angle formed with the side is close to 180 degrees (that is, when tracking the side, from the target side) It is the edge that should be traced to the next edge, and the edge that is considered to be the most naturally connected handwritten). For example, the one closest to 180 degrees with the side ac is the side dc of 175 degrees, and this is stored in the side column 96 that is the selected set corresponding to the side ac in the side information table 90.

FIG. 10 is a conceptual module configuration diagram of a configuration example of the shape analysis extraction module 13.
The shape analysis extraction module 13 includes a graph model construction module 101, a closed curve detection module 102, a graphic determination module 103, and a duplicate graphic removal module 104.
The graph model construction module 101 is connected to the closed curve detection module 102 and is equivalent to the graph model construction module 34 described above.

  The closed curve detection module 102 is connected to the graph model construction module 101 and the figure determination module 103, and by analyzing the graph model generated by the graph model construction module 101, a closed figure (closed curve, plural curves) is generated. Detecting a closed figure formed by a line segment). The detection of the closed figure here uses the vertex information table 80 to detect and detect all possible points as to whether or not each point constitutes a closed curve. In other words, a path starting from the start point and returning to the start point is obtained by full search of the graph.

  The graphic determination module 103 is connected to the closed curve detection module 102 and the duplicate graphic removal module 104, and determines the shape of the closed graphic. That is, the graphic determination module 103 recognizes the shape of the closed graphic after performing polygon approximation on the closed graphic detected by the closed curve detection module 102. The polygon approximation process does not necessarily have to be performed. The processing of the graphic determination module 103 will be described in the processing from step S1106 to step S1118 of the flowchart shown in FIG.

  The duplicate graphic removal module 104 is connected to the graphic judgment module 103, and deletes duplicate graphics from the closed graphics extracted by the graphic judgment module 103. Note that the processing by the duplicate graphic removal module 104 is not necessarily performed.

The process of the duplicate graphic removal module 104 will be described with reference to FIG.
FIG. 12A shows a line figure 1201 that has been thinned. FIG. 12B shows a line figure (circle) 1202, a line figure (triangle (small)) 1203, and a line figure (triangle (large)) 1204 separated by the figure determination module 103. The line figure 1201 shown in FIG. 12A extracts all possible closed figures even though the circle and the triangle are in contact with each other, and thus, the line figure (triangle (small)) is extracted in this way. ) 1203 is also extracted.

An example of processing by the duplicate graphic removal module 104 is as follows.
(1) Sort closed figures in descending order of the number of pixels. For example, in the case of FIG. 12B, a line figure (triangle (large)) 1204, a line figure (circle) 1202, and a line figure (triangle (small)) 1203 are obtained.
(2) Check the number of pixels overlapping with other figures in order, and delete the closed figure if the ratio of overlapping parts is large. The determination as to whether or not the ratio is large is based on comparison with a threshold value (a predetermined value, an average of the overlapping rates of other closed figures, etc.). For example, as shown in FIG. 12C, the line figure (triangle (small)) 1203 overlaps with the line figure (triangle (large)) 1204 in the two sides of the triangle. (The thick line portion in FIG. 12C) is large. At this time, one of the overlapping figures is deleted, but the one with the lower sort order (the one with the smaller total number of pixels) is selected. Therefore, a line figure (triangle (small)) 1203 is a deletion target.

FIG. 11 is a flowchart illustrating a processing example performed by the shape analysis extraction module 13.
In step S1102, the graph model construction module 101 generates a graph model in the same manner as the graph model construction module 34 described above.
In step S1104, the closed curve detection module 102 detects a closed figure as described above.

In step S1106, the graphic determination module 103 performs processing for representing a curve with a plurality of straight lines. For example, a method of detecting corner points and connecting them with straight lines is used.
In step S1108, the graphic determination module 103 determines whether or not the number of lines of the closed graphic is three. If it is 3, the process proceeds to step S1118; otherwise, the process proceeds to step S1110.
In step S1110, the graphic determination module 103 proceeds to step S1118 if the number of line segments of the closed graphic is four and all the angles are close to 90 degrees, and proceeds to step S1112 otherwise. .

In step S <b> 1112, the graphic determination module 103 removes jaggies (jagged lines).
In step S1114, the graphic determination module 103 performs polygon approximation processing on the graphic from which jaggy has been removed. Processing similar to that in step S1106 may be performed, or processing may be performed by changing the threshold value.
Further, as a polygon approximation process, a process using a two-division method may be performed. An example of processing by the two-division method is as follows (see “Digital Image Processing” published by CG-ARTS Association, Masatoshi Okutomi).
(1) The distance is calculated from each pixel with respect to the line segment connecting the start point and the end point of the line segment sequence.
(2) If the maximum distance is less than a certain threshold (predetermined, average distance, etc.), that line segment is adopted. Otherwise, the pixel column is divided by the pixel, and the process (1) is repeated recursively on the divided pixel column.
(3) If the pixel column is not divided, the process is terminated.

In step S1116, the graphic determination module 103 proceeds to step S1118 if the number of line segments of the closed graphic is 6 or more and the number of corners (for example, corners of 100 degrees or less) is 2 or less, otherwise In the case of, the process ends (step S1122).
In step S1118, the graphic determination module 103 determines that the closed graphic is a triangle if the process proceeds from step S1108, and determines that the closed graphic is a rectangle if the process proceeds from step S1110. When proceeding from, it is determined that the closed figure is a circle. Then, the closed figure is extracted.
In step S1120, the overlapping graphic removal module 104 deletes the overlapping closed graphic as described above.
In this flowchart, the determination of a triangle, a rectangle, and a circle is performed, but the determination may be made using other graphic features.

FIG. 13 is a flowchart illustrating a processing example when the shape or number of line figures included in the input image is known. That is, the process of the post-processing module 14 will be mainly described. The post-processing module 14 performs the processing from step S1308 to step S1320.
In step S1302, the preprocessing module 11 performs preprocessing (noise removal processing and thinning processing) as described above.
In step S1304, the line tracking extraction module 12 performs line tracking extraction processing (open graphic extraction processing) as described above.
In step S1306, the shape analysis extraction module 13 performs shape analysis extraction processing (closed graphic extraction processing) as described above.
In step S1308, the total number of figures extracted in steps S1304 and S1306 is compared with a preset number of figures. If they are different, the process proceeds to step S1310. Otherwise, the process proceeds to step S1320. Furthermore, the figure shape extracted in step S1304 and step S1306 is compared with a preset figure shape. If they are different, the process proceeds to step S1310. Otherwise, the process proceeds to step S1320.

In step S1310, it is determined whether there is a contact node candidate. If there is, the process proceeds to step S1312, and if not, the process proceeds to step S1318. Such a determination is made because there are many cases where the processing is erroneous in the contact node. Whether or not the node is a contact node is determined by an angle. That is, the determination is made based on whether there are two angles having an angle close to 180 degrees.
Here, the contact node will be described with reference to FIG. As shown in the figure, in the case of a figure in which two circles are in contact with each other, the intersection path shown in the figure is divided into four line segments and extracted as a pair of line segments (step S710 in the flowchart shown in FIG. 7). 1402 is selected. This is because the angle formed by these two line segments is close to 180 degrees. However, since this figure is originally in contact with two circles, a pair of sides should be extracted so that a contact path 1401 is obtained.

In step S1312, re-analysis is performed at the node targeted in step S1310 to search for a pair that can be contacted. Here, the search target is set to the second candidate close to 180 degrees. In the process of step S1312, the side field 96 that is the selected set in the side information table 90 is changed.
In step S1314, the processes of steps S1304 and S1306 (that is, the processes of the line tracking extraction module 12 and the shape analysis extraction module 13) are executed again.
In step S1316, the process of step S1308 is performed. That is, it is determined whether the re-execution result is correct. If they are still different, the process proceeds to step S1318, and if they match, the process proceeds to step S1320.
In step S1318, even when there is no contact node or when it is executed again, an error report indicating that the processing could not be performed is made because the number or shape of the preset figure is different. .
In step S1320, other post-processing is performed. For example, the line width is restored, the missing portion is interpolated, the data is processed so that the processing result according to the present embodiment can be used by another apparatus, and the like.

  FIG. 15 is a flowchart illustrating a processing example when the number of line figures included in the input image is known. In the flowchart shown in FIG. 13, the number and shape are compared with preset values (steps S1308 and S1316), but in FIG. 15, only the number of line figures is the comparison target. In other words, steps S1508 and S1516 are different from steps S1308 and S1316, respectively, and only the number of line figures is to be compared. The processing of each other step is the same as each step in the flowchart shown in FIG.

  FIG. 16 is a flowchart illustrating a processing example when the processing by the post-processing module 14 is not performed. That is, it is a case where step S1308 and subsequent steps in the flowchart shown in FIG. 13 are not performed. The processing from step S1602 to step S1606 is the same as the processing from step S1302 to step S1306.

  A hardware configuration example of the present embodiment will be described with reference to FIG. The configuration shown in FIG. 17 is configured by a personal computer (PC), for example, and shows a hardware configuration example including a data reading unit 1717 such as a scanner and a data output unit 1718 such as a printer.

  A CPU (Central Processing Unit) 1701 executes various modules described in the above-described embodiment, that is, the modules such as the pre-processing module 11, the line tracking extraction module 12, the shape analysis extraction module 13, and the post-processing module 14. It is a control part which performs the process according to the computer program which described the sequence.

  A ROM (Read Only Memory) 1702 stores programs used by the CPU 1701, calculation parameters, and the like. A RAM (Random Access Memory) 1703 stores programs used in the execution of the CPU 1701, parameters that change as appropriate in the execution, and the like. These are connected to each other by a host bus 1704 including a CPU bus.

  The host bus 1704 is connected to an external bus 1706 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 1705.

  A keyboard 1708 and a pointing device 1709 such as a mouse are input devices operated by an operator. The display 1710 includes a liquid crystal display device or a CRT (Cathode Ray Tube), and displays various types of information as text and image information.

  An HDD (Hard Disk Drive) 1711 has a built-in hard disk, drives the hard disk, and records or reproduces a program executed by the CPU 1701 and information. The hard disk stores received images, vertex information table 80, edge information table 90, and the like. Further, various computer programs such as various other data processing programs are stored.

  The drive 1712 reads out data or a program recorded on a removable recording medium 1713 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, and the data or program is read out as an interface 1707 and an external bus 1706. , The bridge 1705, and the RAM 1703 connected via the host bus 1704. The removable recording medium 1713 can also be used as a data recording area similar to the hard disk.

  The connection port 1714 is a port for connecting the external connection device 1715 and has a connection unit such as USB or IEEE1394. The connection port 1714 is connected to the CPU 1701 and the like via an interface 1707, an external bus 1706, a bridge 1705, a host bus 1704, and the like. A communication unit 1716 is connected to the network and executes data communication processing with the outside. The data reading unit 1717 is a scanner, for example, and executes document reading processing. The data output unit 1718 is a printer, for example, and executes document data output processing.

  Note that the hardware configuration illustrated in FIG. 17 illustrates one configuration example, and the present embodiment is not limited to the configuration illustrated in FIG. 17, and is a configuration capable of executing the modules described in the present embodiment. I just need it. For example, some modules may be configured with dedicated hardware (for example, Application Specific Integrated Circuit (ASIC), etc.), and some modules are in an external system and connected via a communication line Alternatively, a plurality of the systems shown in FIG. 17 may be connected to each other via communication lines so as to cooperate with each other. Further, it may be incorporated in a copying machine, a fax machine, a scanner, a printer, a multifunction machine (an image processing apparatus having any two or more functions of a scanner, a printer, a copying machine, a fax machine, etc.).

  In the above embodiment, the data structures shown in FIGS. 8 and 9 are not limited to these data structures, and may be other data structures. For example, the table structure may be a link structure or the like. Further, the data items are not limited to those shown in the drawings, and may have other data items (for example, line segment length).

The program described above may be provided by being stored in a recording medium, or the program may be provided by communication means. In that case, for example, the above-described program may be regarded as an invention of a “computer-readable recording medium recording the program”.
The “computer-readable recording medium on which a program is recorded” refers to a computer-readable recording medium on which a program is recorded, which is used for program installation, execution, program distribution, and the like.
The recording medium is, for example, a digital versatile disc (DVD), which is a standard established by the DVD Forum, such as “DVD-R, DVD-RW, DVD-RAM,” and DVD + RW. Standards such as “DVD + R, DVD + RW, etc.”, compact discs (CDs), read-only memory (CD-ROM), CD recordable (CD-R), CD rewritable (CD-RW), etc. MO), flexible disk (FD), magnetic tape, hard disk, read only memory (ROM), electrically erasable and rewritable read only memory (EEPROM), flash memory, random access memory (RAM), etc. It is.
The program or a part of the program may be recorded on the recording medium for storage or distribution. Also, by communication, for example, a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a wired network used for the Internet, an intranet, an extranet, or a wireless communication It may be transmitted using a transmission medium such as a network or a combination of these, or may be carried on a carrier wave.
Furthermore, the program may be a part of another program, or may be recorded on a recording medium together with a separate program. Moreover, it may be divided and recorded on a plurality of recording media. Further, it may be recorded in any manner as long as it can be restored, such as compression or encryption.

It is a conceptual module block diagram about the structural example of this Embodiment. It is explanatory drawing which shows the whole process including this Embodiment. It is a conceptual module block diagram about the structural example of a line tracking extraction module. It is a flowchart which shows the process example which a line tracking extraction module performs. It is explanatory drawing which shows the process example which a false line segment extraction module performs. It is explanatory drawing which shows the process example which a graph model construction module and a node analysis module perform. It is a flowchart which shows the process example which a graph model construction module and a node analysis module perform. It is explanatory drawing which shows the data structure of a vertex information table. It is explanatory drawing which shows the data structure of a side information table. It is a conceptual module block diagram about the structural example of a shape analysis extraction module. It is a flowchart which shows the process example which a shape analysis extraction module performs. It is explanatory drawing which shows the process example which an overlapping figure removal module performs. It is a flowchart which shows the process example in case the shape or number of the line figure contained in an input image is known. It is explanatory drawing which shows the example of the contact of a line figure. It is a flowchart which shows the process example in case the number of the line figures contained in an input image is known. It is a flowchart which shows the process example when the process by a post-processing module is not performed. It is a block diagram which shows the hardware structural example of the computer which implement | achieves this Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Pre-processing module 12 ... Line tracking extraction module 13 ... Shape analysis extraction module 14 ... Post-processing module 21 ... Accepted image 22 ... Additional recording extraction module 23 ... Additional recording image 24 ... Single stroke line figure separation module 25 ... Single stroke line segment image 26 ... One-stroke line image 27 ... One-stroke line image 31 ... Pre-processing module 32 ... Feature point / line segment extraction module 33 ... False line segment extraction module 34 ... Graph model construction module 35 ... Node analysis module 36 ... Open figure extraction module 101 ... Graph model construction module 102 ... Closed curve detection module 103 ... Figure judgment module 104 ... Duplicate figure removal module

Claims (8)

Extraction means for extracting features and line segments related to points in the image including line figures;
By said thus extracted feature points to the extraction means and vertices in the graph model, the edges in the graph model line segment extracted by the extraction means, a graph model generating means for generating a graph model,
An open figure extraction means for extracting an open figure having an end point by analyzing the graph model generated by the graph model generation means;
Analyzing the graph model generated by the graph model generating means, and comprising a closed figure extracting means for extracting a closed figure having no end points ;
The image processing apparatus according to claim 1, wherein the closed figure extracting means deletes duplicate figures from the extracted closed figures . Extraction means for extracting features and line segments related to points in the image including line figures;
By said thus extracted feature points to the extraction means and vertices in the graph model, the edges in the graph model line segment extracted by the extraction means, a graph model generating means for generating a graph model,
An open figure extraction means for extracting an open figure having an end point by analyzing the graph model generated by the graph model generation means;
Analyzing the graph model generated by the graph model generating means to extract a closed figure extracting means having no end points ; and
It is determined whether or not the total of the open graphic extracted by the open graphic extraction means and the closed graphic extracted by the closed graphic extraction means is different from a predetermined number of graphics included in the image. Judgment means,
If it is determined by the determination means, the graph model is changed according to the location where the line segment is in contact with the image, and the processing by the open graphic extraction means or the closed graphic extraction means is performed. An image processing apparatus comprising: a reprocessing unit that performs the processing. Extraction means for extracting features and line segments related to points in the image including line figures;
By said thus extracted feature points to the extraction means and vertices in the graph model, the edges in the graph model line segment extracted by the extraction means, a graph model generating means for generating a graph model,
An open figure extraction means for extracting an open figure having an end point by analyzing the graph model generated by the graph model generation means;
Analyzing the graph model generated by the graph model generating means to extract a closed figure extracting means having no end points ; and
It is determined whether or not the shape of the open graphic extracted by the open graphic extraction means or the shape of the closed graphic extracted by the closed graphic extraction means is different from a predetermined graphic shape included in the image. Judgment means,
If it is determined by the determination means, the graph model is changed according to the location where the line segment is in contact with the image, and the processing by the open graphic extraction means or the closed graphic extraction means is performed. An image processing apparatus comprising: a reprocessing unit that performs the processing. Based on the feature points and line segments extracted by the extraction means, comprising a false line segment identifying means for identifying a false line segment out of the line segments,
The image processing apparatus according to any one of claims 1 to 3, wherein the graph model generation unit generates a graph model based on results obtained by the extraction unit and the false line segment identification unit. 4. The closed figure extracting means extracts the closed figure by analyzing the graph model, and after recognizing the polygon, recognizes the shape of the closed figure . The image processing apparatus according to item 1 . Computer
Extraction means for extracting features and line segments related to points in the image including line figures;
By said thus extracted feature points to the extraction means and vertices in the graph model, the edges in the graph model line segment extracted by the extraction means, a graph model generating means for generating a graph model,
An open figure extraction means for extracting an open figure having an end point by analyzing the graph model generated by the graph model generation means;
By analyzing the graph model generated by the graph model generating means, it functions as a closed figure extracting means for extracting a closed figure having no end points ,
The image processing program characterized in that the closed figure extracting means deletes duplicate figures from the extracted closed figures . Computer
Extraction means for extracting features and line segments related to points in the image including line figures;
By said thus extracted feature points to the extraction means and vertices in the graph model, the edges in the graph model line segment extracted by the extraction means, a graph model generating means for generating a graph model,
An open figure extraction means for extracting an open figure having an end point by analyzing the graph model generated by the graph model generation means;
Analyzing the graph model generated by the graph model generating means to extract a closed figure extracting means having no end points ; and
It is determined whether or not the total of the open graphic extracted by the open graphic extraction means and the closed graphic extracted by the closed graphic extraction means is different from a predetermined number of graphics included in the image. Judgment means,
If it is determined by the determination means, the graph model is changed according to the location where the line segment is in contact with the image, and the processing by the open graphic extraction means or the closed graphic extraction means is performed. An image processing program that functions as a reprocessing unit that performs the processing. Computer
Extraction means for extracting features and line segments related to points in the image including line figures;
By said thus extracted feature points to the extraction means and vertices in the graph model, the edges in the graph model line segment extracted by the extraction means, a graph model generating means for generating a graph model,
An open figure extraction means for extracting an open figure having an end point by analyzing the graph model generated by the graph model generation means;
Analyzing the graph model generated by the graph model generating means to extract a closed figure extracting means having no end points ; and
It is determined whether or not the shape of the open graphic extracted by the open graphic extraction means or the shape of the closed graphic extracted by the closed graphic extraction means is different from a predetermined graphic shape included in the image. Judgment means,
If it is determined by the determination means, the graph model is changed according to the location where the line segment is in contact with the image, and the processing by the open graphic extraction means or the closed graphic extraction means is performed. An image processing program that functions as a reprocessing unit that performs the processing.

Images