CN102456139B - Image processing equipment and image processing method - Google Patents

Abstract

The present invention provides an image processing device and an image processing method. An image processing device includes a line information receiving unit, a prediction determining unit, a feature amount calculating unit, and a line determining unit. The line information receiving unit receives a set of information indicating (i) information about an image which may be a line; and (ii) line elements which are rectangular blocks of pixels constituting the line. The prediction determination unit determines whether the target line element matches the predicted value based on the received information. The predicted value indicates the predicted line element when the target line element constitutes the line. The feature quantity calculation unit calculates the feature quantity of the image when the prediction determination unit determines that the target line element does not match the predicted value. The line determining unit determines whether the image is a line based on the calculated feature amount.

Description

Image processing device and image processing method

technical field

The present invention relates to an image processing device and an image processing method.

Background technique

Techniques for extracting line segments from images are known in the art.

As one of these techniques, JP-A-08-016782 discloses a line segment extraction system that can extract a line segment from a binary drawing image even when the line segment is inclined, slightly curved in the middle, or disconnected in the middle . A run data extraction unit extracts a run data group by scanning the binary drawing image longitudinally and laterally. The indication point detection unit detects the positions of the start indication point and the end indication point. The direction data selection unit selects one of the longitudinal and lateral direction data sets according to the inclination between the start indication point and the end indication point. The starting trend detection unit detects trend data whose trend length is shorter than a specified value and whose distance from the start indication point is the smallest from the selected trend data group, and defines the detected trend data as the starting trend. The trend tracking unit tracks the trend data linked from the start trend to the end indication point until the position of the end indication point, and defines the link of the obtained trend data as a line segment.

Contents of the invention

An object of the present invention is to provide an image processing apparatus and an image processing method that can prevent an image that is not a line from being erroneously determined as a line.

[1] According to an aspect of the present invention, an image processing apparatus includes a line information receiving unit, a prediction determining unit, a feature amount calculating unit, and a line determining unit. The line information receiving unit receives a set of information indicating (i) information about an image which may be a line; and (ii) line elements which are rectangular blocks of pixels constituting the line. The prediction determining unit determines whether the target line element matches the predicted value of the target line element based on the information indicating the line element received by the line information receiving unit. The prediction value indicates a line element predicted when the target line element constitutes a line at the position of the target line element. The feature amount calculation unit calculates the feature amount of the image when the prediction determination unit determines that the target line element does not match the predicted value. The line determination unit determines whether the image is a line based on the feature quantity calculated by the feature quantity calculation unit.

[2] In the image processing device according to [1], an image receiving unit and a line extracting unit are further included. The image receiving unit receives an image. The line extracting unit extracts images that may be lines from the images received by the image receiving unit, and extracts a set of information indicating line elements of the lines. The line information receiving unit receives a set of information indicating line elements extracted by the line extracting unit.

[3] In the image processing device according to [1] or [2], the prediction determination unit has a plurality of conditions to determine whether the target line element matches the predicted value. The feature quantity calculation unit extracts the feature quantity based on a condition type when the prediction determination unit determines that the target line element does not match the predicted value.

[4] In the image processing device according to [2] or [3], the image receiving unit receives an image including a character string. The line extracting unit extracts images that may be lines from the images received by the image receiving unit based on the direction of the character string.

[5] In the image processing apparatus according to [2] or [3], the image receiving unit receives the image including the frame line. The line extracting unit extracts an image that may be a line from the image received by the image receiving unit based on the direction of the frame line.

[6] According to another aspect of the present invention, an image processing method includes receiving a set of information indicating (i) information about an image that may be a line, (ii) line elements that are rectangular blocks of pixels constituting the line ; Determine whether the target line element matches the predicted value of the target line element based on the received information indicating the line element, and the predicted value indicates the predicted line element when the target line element forms a line at the position of the target line element; when determining When the target line element does not match the predicted value, calculating a feature quantity of the image; and determining whether the image is a line based on the calculated feature quantity.

According to the image processing apparatus of [1], it is possible to prevent an image that is not a line from being erroneously determined as a line.

According to the image processing apparatus of [2], it is possible to prevent an image that is not a line from among images extracted as possible lines from being erroneously determined as a line.

According to the image processing apparatus of [3], whether an image corresponds to a line can be determined according to a feature based on a condition type when a mismatch with a predicted value is determined.

According to the image processing device of [4], a line used with a character string can be extracted.

According to the image processing apparatus of [5], a frame line can be extracted from an image containing the frame line.

According to the image processing method of [6], it is possible to prevent an image that is not a line from being erroneously determined as a line.

Description of drawings

Exemplary embodiments of the present invention will be described in detail below based on the accompanying drawings, in which:

Fig. 1 is a conceptual module construction view according to an example construction of the first embodiment;

FIG. 2 is an explanatory diagram showing an example target image;

FIG. 3 is an explanatory diagram showing an example image read from a scanner or the like;

FIG. 4 is an explanatory diagram showing an example of line extraction;

FIG. 5 is an explanatory diagram showing an example received image;

FIG. 6 is an explanatory diagram showing example line elements extracted from a received image;

FIG. 7 is an explanatory diagram showing an example line element;

FIG. 8 is an explanatory diagram showing example corrected and supplemented line elements;

FIG. 9 is an explanatory diagram showing an example line element spanning a plurality of pixels;

FIG. 10 is an explanatory diagram showing an example of determining and observing failure types;

11 is a flowchart illustrating example processing performed by an employment determination module;

Fig. 12 is a conceptual module construction view according to an example construction of the second embodiment;

Fig. 13 is a conceptual module construction view according to an example construction of the third embodiment;

Fig. 14 is a conceptual module construction view according to an example construction of the fourth embodiment;

Fig. 15 is a conceptual module construction view according to an example construction of the fifth embodiment;

FIG. 16 is a block diagram illustrating an example hardware configuration for computationally implementing this embodiment.

detailed description

Various exemplary embodiments suitable for realizing the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a conceptual block configuration view of an example configuration according to the first embodiment.

A "module" as used herein generally refers to a component such as logically separable software (computer program), hardware, and the like. Therefore, the modules in this embodiment include not only modules in computer programs but also modules in hardware configurations. Therefore, the embodiments describe all computer programs (including programs that cause a computer to perform steps, programs that make a computer an apparatus, and programs that cause a computer to realize functions) that make the embodiments a module, system, and method. For the convenience of description, when the embodiment refers to a computer program, "store", "stored" or its equivalent expression used herein means that the computer program is stored in the storage device or the computer program is controlled to be stored in the storage device . Although there is a one-to-one correspondence between modules and functions, when installed, one module can be constructed as one program, multiple modules can be constructed as one program, or one module can be constructed as multiple programs. Multiple modules can be executed by one computer, or one module can be executed by multiple computers in a distributed or parallel environment. A module can contain other modules. As used herein, the term "connection" includes not only a physical connection but also a logical connection (data transmission, instruction, reference relationship between data, etc.). As used herein, the term "predetermined" refers to a determination before the target process, including not only the determination before the embodiment starts the process, but also the The circumstances and conditions at the moment, or up to that moment.

As used herein, the term "system" or "equipment" not only refers to a plurality of computers, hardware, equipment, etc. interconnected (including one-to-one communication connections) through communication means (such as a network), but also includes a computer, hardware, equipment, etc. In this specification, "apparatus" is synonymous with "system". Of course, "system" does not include any aspect of a society's "structure" (social institution) that is merely a human judgment.

When different modules perform different processing or one module performs different processing, information for the processing is read from the storage unit, and the processing result is written in the storage unit. Therefore, descriptions about reading information from memory cells before processing and writing information into memory cells after processing may be omitted. The storage unit used herein may include a hard disk, a random access memory (RAM), an external storage medium, a storage unit via a communication line, a register within a central processing unit (CPU), and the like.

The image processing apparatus of the first embodiment determines whether a target image corresponds to a line, and includes a line information receiving module 110, an observation failure analysis module 120, a feature quantity calculation module 130, and an adoption determination module 140, as shown in the example of FIG.

First, the target image in the first embodiment will be described.

FIG. 2 is an explanatory diagram showing an example target image. The shown images include line images (patterns) and non-line images.

FIG. 3 is an explanatory diagram showing an example image (pixel data) obtained when a scanner or the like reads the image in FIG. 2 . In this example, the line to be extracted in the target image is not necessarily represented as a single area that is regarded as a straight line, which is due to the influence of noise, and may also be read by a scanner, etc. as a dotted line, dashed line, Non-linear patterns, etc. More specifically, crushed or deficient threads and the like may occur.

FIG. 4 is an explanatory diagram showing an example of line extraction.

In this example, a line is identified as a set of regions corresponding to pixel runs extending in a direction intersecting (eg, perpendicular to) the line direction. Hereinafter, these areas are referred to as "line elements".

A first embodiment involves detecting and merging pixel strings corresponding to lines to be extracted (ie observed line elements), as shown in the example of FIG. 4 . The combined result of the pixel strings corresponds to a line image.

In addition, solid line elements (ie, correction line elements) may be generated based on these pixel strings. For example, in FIG. 4 , observation line elements 411 to 428 indicated by dotted lines correspond to line elements observed within the image, while correction line elements 451 to 468 indicated by gray rectangles correspond to line elements based on observation. 411 and other corrected line elements. For example, when the observation line element 411 is equal to the correction line element 451 , the observation line element 412 moves down to correspond to the correction line element 452 , and the observation line element 414 moves up to correspond to the correction line element 454 . Also, when no line element is observed between the observed line element 422 and the observed line element 425 , a correction line element 463 and a correction line element 464 are added.

Next, line elements will be described.

FIG. 5 is an explanatory diagram showing an example received image. The received image 500 is divided into background pixels 510 and pattern pixels 520 composed of pixel strings 530 and the like. An example of extracting a horizontal line from this image 500 will be described below.

FIG. 6 is an explanatory diagram showing example line elements extracted from a received image. The pattern pixel 520 is made up of a set of line elements 640 and the like. In this example, a line element refers to an area corresponding to a rectangular pixel block that can constitute a line or the aforementioned pixel string. Examples of "information indicating line elements" may include information used to draw line elements, specifically, line element drawing positions, line element sizes, and the like.

FIG. 7 is an explanatory diagram showing an example line element 710 . In this example, a line element refers to an area corresponding to a pixel string extending in a direction (position direction in FIG. 7 ) intersecting the line direction (travel direction in FIG. 7 ).

Suppose a line is composed of multiple line elements, for example, the kth line element has the following information:

t _k : Line element thickness

P _k : line element position

t _k represents information on the shape of the line element, and it corresponds to the length in the direction intersecting with the line direction. That is, t _k is information corresponding to the thickness of the line.

P _k represents position information. For example, for a horizontal line, P _k is information corresponding to the image height.

As shown in the example in FIG. 7 , a line element actually observed from an image has a minimum unit of one pixel. However, the smallest information unit of a line element is not necessarily a pixel, and may be a unit smaller than a pixel. For example, solid line elements may be obtained in units smaller than one pixel based on the obtained line elements. Alternatively, a line element may be generated from a state where no line element is observed, such as line extraction including correction processing for supplementing a line element in a halfway cut portion as shown in FIG. 8 .

FIG. 8 is an explanatory diagram showing an example corrected and supplemented line element. In this figure, the observation line elements 811 to 828 indicated by the hatched rectangles correspond to the observed line elements within the image, while the correction line elements 851 to 868 indicated by the gray rectangles correspond to the line elements based on the observation line element 811 etc. Corrected line elements. For example, when no line element is observed between observed line element 820 and observed line element 822 , correction line element 861 is added.

Additionally, not all of the above information is necessarily used. For example, if the line position is used for line extraction (described later), position information on each line element is unnecessary.

Additionally, if processing a multivalued image, pixel density information can be provided. For example, a line element whose color information changes suddenly may be determined as an observation failure.

Additionally, a line element can be connected to multiple pixels, as shown in Figure 9. For example, an observation line element may correspond to one pixel, while a correction line element may correspond to multiple pixels. In FIG. 9 , observation line elements 811 to 828 indicated by hatched rectangles correspond to observed line elements within the image, while correction line elements 951 to 959 indicated by gray rectangles correspond to lines based on observation line element 811 etc. Corrected line elements. For example, the observation line element pair consisting of observation line element 811 and observation line element 812 is moved up to correspond to correction line element 951, and when there is no observation line element between observation line element 820 and observation line element 822, an additional A correction line element 956 is generated.

The line information receiving module 110 is connected to the observation failure analysis module 120 and the feature quantity calculation module 130 . The line information receiving module 110 receives the line information 108 and transmits the line information 112 to the observation failure analysis module 120 and the feature quantity calculation module 130 . The line information 108 refers to information related to an image that may be a line, and corresponds to a set of information indicating line elements. The first embodiment involves determining whether the line information 108 constitutes a line.

The observation failure analysis module 120 , which is a prediction determination unit, is connected to the line information reception module 110 and the feature quantity calculation module 130 . The observation failure analysis module 120 determines whether the line information 112 is successfully observed or not, and transmits the observation failure information 122 to the feature amount calculation module 130 . If the line element is predicted to constitute a line at the line element's location, the observation failure analysis module 120 determines whether the target line element matches the predicted value of the line element. This determination of whether the target line element matches the predicted value of the line element if the line element is predicted to constitute a line at the location of the line element is also referred to as an "observation". "Observation success" refers to a determination that matches the predicted value, and "observation failure" refers to a determination that does not match the predicted value.

In addition, for the observation failure, the observation failure analysis module 120 may also output the failure type as the observation failure information 122 .

The processing performed by the observation failure analysis module 120 will be described in detail below.

The observation failure analysis module 120 receives the line information 112, and outputs observation failure information 122 (which is information having a type of observation failure).

As shown in FIG. 10 , consider a line composed of line elements extending in the direction of travel. In this example, the following information is obtained from line information 108:

f _t (k): line thickness at position k

f _p (k): line position at position k

f _t (k) and f _p (k) correspond to ideal values for the line element at position k.

In this case, the line element thickness t _k at position k does not have to be equal to f _t (k).

Additionally, line element position p _{k at position k} does not have to be equal to f _p (k).

Also, in some cases there is no line element at position k.

Thus, a line element at position k can have observation failures of the following types:

A) There is no line element;

B) The difference between p _k and f _p (k) is greater than a preset threshold;

C) The difference between t _k and f _t (k) is greater than a preset threshold.

These types of observation failures will be described in detail below using the example of FIG. 10 . In FIG. 10 , gray rectangles represent line elements indicated by line information 112 (observed line elements). A region (line) surrounded by dotted lines at the top and bottom represents an ideal line calculated based on the line information 112, and a dotted line represents the center of the ideal line.

In the example of FIG. 10 , there is no line element at the position next to the third line element from the leftmost (the position indicated by A). Here is an example of observation failure (A).

In the example of FIG. 10 , the seventh line element from the far left (position indicated by B) protrudes upwards, which means that the difference between p _k and f _p (k) is greater than a preset threshold. This is an example of observation failure (B).

The type of observation failure differs depending on whether the difference between t _k and f _t (k) is positive or negative. In the example of FIG. 10 , the eleventh line element from the leftmost (position indicated by C-1 ) is too thick compared to that of the ideal line. In contrast, in the example of FIG. 10 , the fifteenth line element from the leftmost (position indicated by C-2 ) is too thin compared to that of the ideal line.

In order to gracefully handle these types of observation failures, the observation failure analysis module 120 has a number of conditions to determine if there is a match with the predicted value.

The first condition is that there is a line element immediately adjacent to the target line element. This corresponds to type A) above. Failure of this condition causes the first type of observation to fail.

The second condition is that the difference between the position of the observed line element and the position of the predicted line element is less than a threshold. This corresponds to type B) above. If the difference is equal to or greater than the threshold, it causes the second type of observation to fail.

The third condition is that the difference between the width of the observed line element and the width of the predicted line element is less than a threshold. This corresponds to type C) above. If the difference is equal to or greater than the threshold, it causes the third type of observation to fail. If the difference is less than the threshold, it results in a failure of the fourth type of observation.

The threshold above can be set as a constant, an integer multiple of f _p (k), an integer multiple of f _t (k), or an integer multiple of the variation (variance) of line elements.

The observation failure analysis module 120 may output observations with type additions, such as A), B), C)-1, and C)-2.

f _t (k) and f _p (k) can be obtained in existing methods. For example, f _t (k) may be obtained as the mean, median or maximum frequency of the observed line information 112 . f _p (k) can be obtained from the position of the observed line information 112 by a least square error function.

Also, it is not necessary to use all line elements to obtain ft( _k ) and fp( _k ). For example, the line elements for which f _t (k) and f _p (k) are to be determined may refer to neighboring line elements. Adjacent as used herein refers to adjacent line elements existing within a predetermined distance from the target line element. In addition, it may also refer to the remaining line elements other than the line element causing obvious observation failure.

Furthermore, other types of observation failures may provide information on the concentration of line elements in addition to the above-mentioned types of observation failures.

The observation failure analysis module 120 can be divided into two modules.

The first module is the prediction value calculation module. If the line elements form a line from the position of the target line element, it calculates the prediction value of the relevant information of the line elements at the target position based on the relevant information of the line elements received by the line information receiving module 110 .

The second module is a prediction determination processing module, which determines whether the predicted value calculated by the predicted value calculation module matches the related information of the line elements received by the line information receiving module 110 .

In addition, the observed failure analysis module 120 also classifies cases that do not match the predicted values according to type.

The feature quantity calculation module 130 is connected to the line information reception module 110 , the observation failure analysis module 120 and the adoption determination module 140 . The feature amount calculation module 130 calculates the feature amount of the line information 112 corresponding to the observation failure information 122 , and transmits the feature amount 132 to the adoption determination module 140 . The feature amount calculation module 130 calculates the image feature when the failure-to-observation analysis module 120 determines that a mismatch with the predicted value occurs. In addition, the feature amount calculation module 130 may extract features based on the type of condition determined by the observation failure analysis module 120 that does not match the predicted value. Specifically, the feature may be an occurrence rate that does not match the predicted value (observation failure).

The feature amount calculation module 130 refers to the line information 112 and the observation failure information 122 to obtain line feature amounts. For each type of observation failure, or by arranging several types of observation failures, the occurrence rate of observation failure is obtained as a feature quantity.

Specifically, it is obtained as follows:

First, the types of observation failures are arranged according to the following equations (1) and (2). In the equation, n{X} represents the number of occurrences of type X of observation failures. The method of arranging observation failure types is not limited to equations (1) and (2). For example, n{C ₂ } can be added to thick_skip. In addition, it is not necessary to schedule the type of observation failure.

[equation 1]

blank_skip=n{A}+n{B}+n{C ₂ } (1)

A: no line element

B: The thickness of the line element is too thin

C ₂ : The position of the line element is too far from the predicted position of the line element

[equation 2]

thick_skip=n{C ₁ } (2)

C ₁ : The thickness of the line element is too thick

It is assumed that the total number of line elements constituting a line, the occurrence rate of observation failure types, that is, feature quantities are obtained according to the following equations (3) and (4).

[equation 3]

ratio_blank_skip=blank_skip/N (3)

[equation 4]

ratio_thick_skip=thick_skip/N (4)

Here, the type of observation failure may not be used to obtain characteristics different from equations (3) and (4). For example, accumulations of length, thickness, position, gradient, distortion, error of ft( _k ) from _tk (ie, thickness) and error of fp( _k ) from _pk (ie, position) are used. In this case, it is possible to obtain the above-mentioned feature amount in addition to the line element causing observation failure.

The adoption determination module 140 (which is a line determination unit) is connected to the feature quantity calculation module 130 . The adoption determination module 140 determines whether or not a line element is being used as a line based on the feature quantity 132 , and outputs adoption/non-adoption information 142 . Here, examples of output may include storage on a storage medium such as a memory card, transmission to another information processing device, and the like.

The employing determining module 140 determines whether the image constituted by the line information 108 received by the line information receiving module 110 is a line based on the feature quantity 132 calculated by the feature quantity calculating module 130 .

The adoption determination module 140 outputs line adoption/non-adoption information 142 based on the line feature quantity 132 . In this case, it may be determined whether each feature satisfies a predetermined condition. For example, the determination processing shown in the example of FIG. 11 is performed. In FIG. 11 , θ _{ratio_blank_skip} and θ _{ratio_thick_skip} are predetermined values.

In addition, determinants can be generated by learning from feature quantities. In this case, feature quantities of different lines can be used. That is, the processing is configured by a secondary classifier that takes at least the line feature quantity on the observation failure type as input.

FIG. 11 is a flowchart illustrating example processing performed by employing the determination module 140 .

In step S1102, it is determined whether the following equation (5) or (6) is established. If yes, the process proceeds to step S1104, otherwise, the process proceeds to step S1106.

[equation 5]

ratio_blank_skip<θ _{ratio_blank_skip} (5)

[equation]

ratio_thick_skip<θ _{ratio_thick_skip} (6)

In step S1104, "adopt" is output.

In step S1106, "not adopted" is output.

The type of observed failure is characterized by solid lines having less blank_skip and more thick_skip, while dashed lines have less thick_skip and more blank_skip. In addition, it is characterized in that the nonlinear pattern has many thick_skips and blank_skips. Therefore, when extracting lines from an image in which solid lines, incomplete lines, wrinkled lines, line intersections, non-line patterns, etc. pattern.

In addition, if the occurrence rate of observation failure is equal to or less than a predetermined value, it may be determined whether or not a line is taken regardless of subsequent determination, or a threshold value for subsequent determination may be updated. This can result in a high probability of line adoption if the extraction line is successfully observed.

Fig. 12 is a conceptual block configuration view according to an example configuration of the second embodiment.

The configuration of the second embodiment includes an image reception module 1210 , a line extraction module 1220 , an observation failure analysis module 120 , a feature amount calculation module 130 , and an adoption determination module 1240 . The same blocks as those in the first embodiment are denoted by the same reference numerals, and description thereof will not be repeated. For blocks denoted by the same reference numerals, the following description involves addition or replacement of the performance or function of the blocks corresponding to the first embodiment. The same applies to the third and subsequent embodiments.

The image receiving module 1210 is connected to the line extracting module 1220 . Image receiving module 1210 receives image 1208 and transmits image 1212 to line extraction module 1220 .

The line extraction module 1220 is connected to the image reception module 1210 , the observation failure analysis module 120 , the feature quantity calculation module 130 , and the adoption determination module 1240 . The line extraction module 1220 extracts a line by aligning line elements from the image 1212, transmits the line information 1222 to the observation failure analysis module 120, the feature quantity calculation module 130, and the adoption determination module 1240, and in the case where there is no line in the image 1212 An invalid signal 1224 is output. The line extracting module 1220 extracts an image that may be a line from the image 1212 received by the image receiving module 1210, and extracts line information 1222 (which is a set of information representing line elements of a line). The prior art disclosed in JP-A-08-016782 and the like can be used for this line extraction processing.

The observation failure analysis module 120 is connected to the line extraction module 1220 and the feature amount calculation module 130 . The observation failure analysis module 120 receives line information 1222 (which is a set of information representing line elements of a line), makes observation success/failure determination of the line information 1222 , and transmits the observation failure information 122 to the feature quantity calculation module 130 . In this case, for observation failure, the type of observation failure may be output together.

The feature quantity calculation module 130 is connected to the line extraction module 1220 , the observation failure analysis module 120 , and the adoption determination module 1240 . Feature calculation module 130 receives line information 1222 extracted by line extraction module 1220 (it is a set of information representing line elements) and observation failure information 122 from observation failure analysis module 120, calculates line information corresponding to observation failure information 122 1222, and transmit the feature 132 to the adoption determination module 1240. Here, other feature quantities may also be calculated.

The adoption determination module 1240 is connected to the line extraction module 1220 and the feature quantity calculation module 130 . The adoption determination module 1240 determines whether an image is adopted as a line based on the feature amount 132 , and outputs a line/invalid signal 1242 . That is, the information representing the line is output to indicate adoption, and the invalid signal ⊥ is output to indicate not to adopt.

Fig. 13 is a conceptual block configuration view according to an example configuration of the third embodiment.

The configuration of the third embodiment includes an image reception module 1210 , a line extraction module 1320 , an observation failure analysis module 1330 , a feature amount calculation module 130 , and an adoption determination module 1240 . The observation failure analysis module 1330 may be integrated with the line extraction module 1320 . That is, for line extraction, in the process of merging line elements, observation failure analysis can be performed sequentially whenever a line element is obtained. Compared with the example of FIG. 12 of the second embodiment, repeated observation failure analysis is not required, which can lead to an increase in processing speed. In addition, memory efficiency can be improved since unnecessary line elements that cause observation failures are not included.

The image receiving module 1210 is connected to the line extracting module 1320 . Image receiving module 1210 receives image 1208 and transmits image 1212 to line extraction module 1320 .

The line extraction module 1320 is connected to the image reception module 1210 , the observation failure analysis module 1330 , the feature amount calculation module 130 , and the adoption determination module 1240 . The observation failure analysis module 1330 is connected to the line extraction module 1320 . In the process of merging line elements in the line extraction module 1320 , the observation failure analysis module 1330 sequentially executes the above-described observation failure analysis processing every time a line element 1322 is obtained from the image 1212 . In addition, the line extraction module 1320 receives observation failure information 1332 from the observation failure analysis module 1330 . Finally, if there is no line in the image 1212, the line extraction module 1320 outputs an invalid signal 1326, and if there is an image that may be a line, the line information and observation failure information 1324 are transmitted to the feature calculation module 130 and the adoption determination module 1240.

The feature amount calculation module 130 is connected to the line extraction module 1320 and the adoption determination module 1240 . The feature quantity calculation module 130 receives the line information and the observation failure information 1324 from the line extraction module 1320, and calculates the feature quantity of the image.

The adoption determination module 1240 is connected to the line extraction module 1320 and the feature quantity calculation module 130 . The adoption determination module 1240 receives line information and observation failure information 1324 from the line extraction module 1320 and the feature quantity 132 from the feature quantity calculation module 130 , and outputs a line/invalid signal 1242 .

Fig. 14 is a conceptual block configuration view according to an example configuration of the fourth embodiment.

The configuration of the fourth embodiment includes an image reception module 1410 , a character string extraction module 1420 , an image processing module 1430 , a character recognition module 1440 and an output module 1450 .

The image receiving module 1410 is connected to a character string extracting module 1420 . The image receiving module 1410 receives an image including a character string. The image may include underlining, strikethrough, etc.

The character string extraction module 1420 is connected to the image receiving module 1410 and the image processing module 1430 . The character string extracting module 1420 extracts a character string from the image received by the image receiving module 1410, and extracts an image that may be a line based on the direction of the character string. String extraction can use existing techniques. For example, a histogram of black pixels in vertical and horizontal directions can be obtained, and the directions of vertical and horizontal writing can be extracted from the distribution of the histograms. Extract possible line images with the orientation of the string as the line direction.

The image processing module 1430 is connected to the character string extraction module 1420 and the character recognition module 1440, and corresponds to one of the image processing device of the first embodiment, the image processing device of the second embodiment, and the image processing device of the third embodiment . That is, it is determined whether the image (possibly a line) extracted by the character string extraction module 1420 is a line.

The character recognition module 1440 is connected to the image processing module 1430 and the output module 1450 . The character recognition module 1440 erases images determined to be lines by the image processing module 1430 in the image, and recognizes only character images.

The output module 1450 is connected to the character recognition module 1440 and outputs a character recognition result of the character recognition module 1440 .

Fig. 15 is a conceptual block configuration view according to an example configuration of the fifth embodiment.

The configuration of the fifth embodiment includes an image receiving module 1510 , a frame line direction specifying module 1520 , an image processing module 1530 , a business form processing module 1540 and an output module 1550 .

The image receiving module 1510 is connected to the frame line direction specifying module 1520, and receives an image including a frame line, for example, a business form image.

The frame line direction specifying module 1520 is connected to the image receiving module 1510 and the image processing module 1530, and extracts images that may be lines from images received by the image receiving module 1510 based on the direction of the frame line. For example, the frame line direction specifying module 1520 specifies and extracts vertical and horizontal lines (mainly used in business tables), and extracts images that may be lines.

The image processing module 1530 is connected to the frame line direction specifying module 1520 and the business form processing module 1540, and corresponds to the image processing device of the first embodiment, the image processing device of the second embodiment, and the image processing device of the third embodiment One. That is, it is determined whether the image that may be a line extracted by the frame line direction specifying module 1520 is a line.

The business form processing module 1540 is connected to the image processing module 1530 and the output module 1550 . The business form processing module 1540 determines the columns in the business form according to the positions of the line images determined as lines in the image by the image processing module 1530 , and recognizes the character images in the columns, thereby associating the determined columns with the character recognition results.

The output module 1550 is connected to the business form processing module 1540 and outputs the business form processing result of the business form processing module 1540 .

An example of the hardware configuration of the image processing apparatus of this embodiment will now be described with reference to FIG. 16 . The hardware configuration shown in FIG. 16 is configured by, for example, a personal computer (PC) or the like, and includes a data reading unit 1617 such as a scanner or the like, a data output unit 1618 such as a printer or the like.

A central processing unit (CPU) 1601 is a control unit that executes processing according to a computer program that describes various modules (for example, the line information receiving module 110, the observation failure analysis module 120, the feature amount calculation module 130) described in the above embodiments. , use determination module 140, image receiving module 1210, line extraction module 1220, use determination module 1240, line extraction module 1320, observation failure analysis module 1330, character string extraction module 1420, image processing module 1430, character recognition module 1440, frame line Direction designation module 1520, image processing module 1530, business form processing module 1540, etc.) execution sequence.

A read only memory (ROM) 1602 stores programs used by the CPU 1601 , operating parameters, and the like. A random access memory (RAM) 1603 stores programs for execution by the CPU 1601 , parameters appropriately changed for execution, and the like. These memories are interconnected through a host bus (such as a CPU bus, etc.) 1604 .

The host bus 1604 is connected to an external bus 1606 such as a Peripheral Component Interconnect/Interface (PCI) bus or the like through a bridge 1605 .

A pointing (point) device 1609 such as a keyboard 1608, a mouse, etc. is an input device manipulated by an operator. A display 1610 such as a liquid crystal display device, a cathode ray tube (CRT), or the like displays various information as text or image information.

A hard disk drive (HDD) 1611 includes a hard disk, and drives the hard disk to record or reproduce programs or information executed by the CPU 1601 . The hard disk stores target images, information indicating line elements, determination results, and the like. In addition, the hard disk stores various computer programs such as data processing programs.

The drive 1612 reads data or programs recorded in a removable recording medium (such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc.) 1613 mounted thereon, and transmits data through an interface 1607, an external bus 1606, a bridge 1605, and a host bus. 1604 provides the read data or program to RAM 1603 . The removable recording medium 1613 can also be used as a data recording area like a hard disk.

The connection port 1614 is a port connected to an external connection device 1615, and includes connection units such as USB, IEEE 1394, and the like. The connection port 1614 is also connected to the CPU 1601 and the like through the interface 1607, the external bus 1606, the bridge 1605, the host bus 1604, and the like. The communication unit 1616 is connected to a network for data communication with the outside. The data reading unit 1617 is, for example, a scanner for reading documents. The data output unit 1618 is, for example, a printer for outputting document data.

The hardware configuration of the image processing apparatus shown in FIG. 16 is an example of configuration, and this embodiment is not limited to the hardware configuration shown in FIG. 16 , and may have any other configuration as long as it can execute the method described in this embodiment. module. For example, some modules can be constructed as dedicated hardware (for example, ASIC (Application Specific Integrated Circuit) etc.), some modules can be in external systems, and connected by communication links, in addition, a plurality of systems shown in Figure 16 can be connected by Communication links are interconnected for cooperation. In addition, the hardware configuration may be incorporated in a copier, a facsimile, a scanner, a printer, a multifunction copier (an image processing apparatus having two or more functions of a scanner, a printer, a copier, and a facsimile), etc. .

The various embodiments described above may be combined (for example, including adding a module in one embodiment to another embodiment or replacing a module in another embodiment), or the technology described in [Background Art] may be used as each The processing content of the module.

The above-mentioned program may be stored in a recording medium, or may be provided by means of communication. In this case, for example, the above-mentioned program can be understood as an invention of "the computer-readable recording medium on which the program is recorded".

The "computer-readable recording medium on which the program is recorded" refers to a computer-readable recording medium on which the program is recorded for installation, execution, distribution, and the like of the program.

For example, recording media may include media such as "DVR-R, DVD-RW, DVD-RAM, etc." (which are standards specified by the DVD Forum) and "DVD+R, DVD+RW, etc." ) such as Digital Versatile Disc (DVD), Compact Disc (CD) such as Read Only Memory (CD-ROM), Recordable CD (CD-R), Rewritable CD (CD-RW), etc., Blu-ray disc (registered trademark), magneto-optical disk (MO), floppy disk (FD), magnetic tape, hard disk, read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM (registered trademark)), flash memory, random access Memory (RAM), etc.

The program or a part of the program can be recorded in a recording medium for storage and distribution. In addition, the program or a part of the program may be transmitted by means of communication, for example, through a network such as a network for a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), the Internet, an intranet, an extranet, etc., or a combination thereof. A transmission medium such as a wired network or a wireless network, or a carrier may be used to carry the data.

The program may be a part of another program, or may be recorded in a recording medium together with a different program. In addition, the program may be divided and recorded in a plurality of recording media. In addition, the program may be recorded in any form (including compression, encryption, etc.) as long as it can be reproduced.

The foregoing description of example embodiments of the present invention has been presented for purposes of illustration and description only, and is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Many changes and modifications will be apparent to those skilled in the art. The selection and description of the embodiments are to better illustrate the principle of the present invention and its practical application, so that those skilled in the art can understand the present invention of various embodiments and various modifications suitable for specific purposes. The scope of the invention is defined by the appended claims and their equivalents.

Claims (6)

1. an image processing equipment, including:

Line information receiving unit, for receiving one group of information of instruction following items: (i) may be for the relevant letter of the image of line Breath and (ii) are as the line element of rectangular block of pixels of composition line；

Prediction determines unit, for determining mesh based on the information indicating described line element received by line information receiving unit Whether graticule element mates with the predictive value of described score element, and described predictive value indicates described score element at described mesh The line element that the position of graticule element is predicted when constituting line；

In described prediction, feature amount calculation unit, for determining that unit does not determines described score element and described predictive value not Timing, calculates the characteristic quantity of described image；And

Line determines unit, and the characteristic quantity for being calculated based on described feature amount calculation unit determines that whether described image is Line,

Wherein, described characteristic quantity is described score element and the unmatched incidence rate of described predictive value, and described characteristic quantity is By there is no the number of times of score element, the thickness of score element and the difference of the thickness of the line element predicted more than predetermined thickness Distance between the degree number of times of threshold value, the position of score element and the position of line element predicted is more than precalculated position threshold value Number of times, the sum of score element determines.

Image processing equipment the most according to claim 1, also includes:

Image receiving unit, is used for receiving described image；And

Straight line extraction unit, may be the image of line for extraction from the image received by described image receiving unit, and Extract one group of information of the line element indicating described line,

Wherein, described line information receiving unit receives described one group of letter of the indicatrix element extracted by described straight line extraction unit Breath.

Image processing equipment the most according to claim 1 and 2,

Wherein, described prediction determines that unit has multiple condition to determine that whether described score element and described predictive value Join, and

Wherein, based on described prediction, described feature amount calculation unit determines that unit determines described score element and described predictive value Condition types when not mating is to extract characteristic quantity.

Image processing equipment the most according to claim 2,

Wherein, described image receiving unit receives the image including character string, and

Wherein, described straight line extraction unit direction based on described character string carries from the image that described image receiving unit receives Taking may be for the image of line.

Image processing equipment the most according to claim 2,

Wherein, described image receiving unit receives the image including wire, and

Wherein, described straight line extraction unit direction based on described wire is extracted from the image that described image receiving unit receives It may be the image of line.

6. an image processing method, including:

Receive instruction following items one group of information, (i) may be the image of line for information about, and (ii) as composition line The line element of rectangular block of pixels；

The prediction of score element and described score element is determined based on the received information indicating described line element Whether value mates, and described predictive value indicates described score element to be predicted when constituting line in the position of described score element Line element；

When determining that described score element does not mates with described predictive value, calculate the characteristic quantity of described image；And

Determine whether described image is line based on the characteristic quantity calculated,

Wherein, described characteristic quantity is described score element and the unmatched incidence rate of described predictive value, and described characteristic quantity is By there is no the number of times of score element, the thickness of score element and the difference of the thickness of the line element predicted more than predetermined thickness Distance between the degree number of times of threshold value and the position of score element and the position of line element predicted is more than precalculated position threshold The number of times of value, the sum of score element determine.