JP2005157905A - Image processing device and method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing device where paper documents or difficult-to-reuse image information can be treated as reusable electronic files. <P>SOLUTION: A composite machine 100 reads image data from a document, and, if pointer information is added to the image data read, obtains an original electronic file according to the pointer information; if no pointer information is added to the image data read, it converts the image data into vector data. During this conversion to the vector data, the content to be converted is varied depending on a file format designated by the user, e.g., word processor format, tabular format, presentation format, and the like. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus, an image processing method, and a program for inputting image data.

  In recent years, paperless offices are rapidly becoming paperless as environmental issues are screamed. Conventionally, a document management system has been constructed by reading a paper document stored in a binder or the like with a scanner and storing it in an image storage device as a database (see, for example, Patent Document 1).

Conventionally, in a multifunction machine such as a copying machine with an extended function, when an image is recorded in advance as a paper document, annotation information indicating the location where the original electronic file of the image is stored is stored in the image storage device. This information is recorded as additional information on the front cover of the document or in the document, and when the document is reused for copying again, the storage location of the original electronic file is detected from the annotation information, and the information on the electronic file is stored. By using it for editing and printing, saving of the entire paper document is reduced (for example, refer to Patent Document 2).
Japanese Patent Laid-Open No. 11-232155 JP 2001-67349 A

  However, the above-described conventional image processing apparatus has the following problems, and improvements have been demanded. That is, in the former document management system, a file for storing a paper document is image information, and therefore some objects of this document cannot be reused. Therefore, when reusing, a figure, a table, and the like have to be newly created using application software.

  In the latter multifunction device, if the original electronic file corresponding to the output paper document can be directly accessed, it can be easily reused. For example, when the original electronic file cannot be accessed, or to the original electronic file in the first place. Documents that do not have the annotation information cannot be reused.

  SUMMARY An advantage of some aspects of the invention is that it provides an image processing apparatus, an image processing method, and a program that enable a paper document or image information that is difficult to be reused to be handled as a reusable electronic file.

  In order to achieve the above object, an image processing apparatus according to the present invention includes vectorization means for generating vector data from image data input from input means for inputting image data, and the generated vector data as a predetermined file. Output means for outputting in a format.

  The image processing method of the present invention includes an input step for inputting image data, a vectorization step for generating vector data from the input image data, and an output step for outputting the generated vector data in a predetermined file format. And have.

  The program of the present invention is a program executed by a CPU in an image processing apparatus, and includes an input step for inputting image data, a vectorization step for generating vector data from the input image data, and the generation step. And outputting the vector data in a predetermined file format.

  According to the present invention, vector data is generated from input image data and output in a predetermined file format, so that a paper document or image information that is difficult to reuse can be handled as a reusable electronic file. be able to. This is equivalent to getting the original electronic file. Moreover, vector data can be generated in a file format suitable for a user-desired application.

  Embodiments of an image processing apparatus, an image processing method, and a program according to the present invention will be described with reference to the drawings. The image processing apparatus of this embodiment is applied to an image processing system.

[First Embodiment]
FIG. 1 is a diagram illustrating a configuration of an image processing system according to the first embodiment. The image processing system according to this embodiment is realized in an environment in which a network (LAN) 107 in the office 10 and a network (LAN) 108 in the office 20 are connected via the Internet 104.

  A LAN 107 constructed in the office 10 is connected to a multifunction device 100, a management PC 101 that controls the multifunction device 100, a client PC 102, a document management server 106, a database 105, and a proxy server 103. On the other hand, a document management server 116, a database 115, and a proxy server 113 are connected to the LAN 108 built in the office 20. The LAN 107 and the LAN 108 are connected to the Internet 104 via proxy servers 103 and 113.

  The multi-function device 100 reads an image of a paper document, performs image processing on the read image signal (preprocessing), and transmits the image data to the management PC 101 via the LAN 109. The management PC 101 is a normal PC and includes an image storage unit, an image processing unit, a display unit, and an input unit. A part of the management PC 101 is configured integrally with the multifunction device 100.

  FIG. 2 is a diagram illustrating an electrical configuration of the multifunction peripheral 100. The multifunction device 100 includes an image reading unit 210, a storage unit 211, a recording unit 212, an input unit 213, a display unit 216, a data processing unit 215, and network interfaces (I / F) 214 and 217. The input unit 213 is provided with a key operation unit.

  The image reading unit 210 includes an auto document feeder (ADF), irradiates a conveyed document with a light source, forms an image of reflected light from the document on a solid-state image sensor with a lens, and outputs a raster image signal from the solid-state image sensor. Is acquired as image information having a density of 600 dpi. When normal copying is performed, the data processing unit 215 performs image processing of the raster image signal to obtain a recording signal (recording data). In the case of a plurality of copies, the recording data for one page is temporarily stored and held in the storage unit 211 and then sequentially output to the recording unit 212 to form an image on a sheet.

  Also, when print data output from the client PC 102 is read into the data processing unit 215 from the LAN 107 via the network IF 214, the print data is converted into raster data that can be recorded (printed) by the data processing unit 215, and then recorded. An image is formed on the paper by the unit 212. An operator's instruction to the multifunction device 100 is issued from a key operation unit provided in the multifunction device 100, a keyboard and a mouse provided in the management PC 101. Further, the display of the input state and the image data being processed are performed on the display unit 216. These series of operations are controlled by a control unit (CPU) in the data processing unit 215.

  The storage unit 211 is also controlled from the management PC 101. Data exchange and control between the MFP 100 and the management PC 101 are performed using a network IF 217 and a directly connected LAN 109.

  The operation of the image processing system having the above configuration will be described. 3 and 4 are flowcharts showing the operation processing procedure of the multifunction peripheral 100. FIG. This processing program is stored in a storage medium in the multifunction peripheral 100 and the management PC 101, and is executed by the CPU in the data processing unit 215 and the CPU in the management PC 101. First, the image reading unit 210 of the multi-function device 100 is operated to scan one original by the raster method, and perform processing (image information input processing) for obtaining image data having a resolution of 600 dpi and 8 bits (step S1). . At this time, the data processing unit 215 performs preprocessing of the acquired image data, and stores image data for one page in the storage unit 211.

  The CPU of the management PC 101 performs block selection (BS) processing (step S2), and separates the area into character / line image portions and halftone image portions for the image data stored in the storage unit 211. Further, among the character / line drawing parts, the character part is separated into blocks each grouped as a lump for each paragraph, the line drawing part is separated into a table and a figure composed of lines, and each is segmented. In addition, the image portion expressed in halftone is divided into so-called independent objects for each block, such as an image portion and a background portion of a block separated into rectangles.

  At this time, a two-dimensional barcode recorded as additional information in the document image or an object corresponding to the URL is detected. In the case of the two-dimensional barcode, the mark is decoded, and in the case of the URL, the character is recognized by OCR. Processing (OCR / OMR processing) is performed (step S3). Then, the pointer information in the storage device storing the original electronic file of the document is detected (step S4). In addition, as a method of adding pointer information, there may be a so-called digital watermark method that is not directly visualized, such as a method of embedding information between characters and a method of embedding in a halftone image.

  It is determined whether pointer information is detected (step S5). When the pointer information is detected, the original electronic file is searched from the address included in the pointer information (step S6A), and the presence or absence of the electronic file is confirmed (step S6). Here, the electronic file is stored in the hard disk of the client PC 102, in the databases 105 and 115 managed by the document management servers 106 and 116 connected to the LANs of the offices 10 and 20, or in the storage unit 211 of the MFP 100. And is searched based on the pointer information obtained in step S3.

  If the electronic file is not found in step S6, or if the electronic file is found but is a so-called image file represented by the PDF format or the tiff format, the process proceeds to step S7. If no pointer information exists in step S5, the process proceeds to step S7.

  Then, a document search process is performed (step S7). In this document search process, a full-text search for extracting words based on the result of OCR performed for each character block is performed, or a layout search is performed based on the array of each object and the attributes of each object.

  As a result of full text or layout search When an electronic file with a high degree of similarity is found, a file candidate is displayed as a thumbnail image, etc., and when an operator needs to be selected from a plurality of file candidates, A file is specified (step S8). Even if the file candidate is only one file, in the present embodiment, the process directly proceeds to step S8. In such a case, the process proceeds to step S9 and S10 without performing step S8. Alternatively, the file candidate address may be notified to the user.

  Then, it is determined whether or not an electronic file is found (step S9). If found, the address of the electronic file specified in step S8 is notified to the user (step S10). If an electronic file is found based on the pointer information in step S6, the user is notified of the address where the electronic file is stored. On the other hand, if the electronic file is not found in step S9, the process proceeds to step S13.

  An electronic file found based on the address is obtained (obtained) (step S11). It is determined whether or not the acquired electronic file is a so-called image file represented by a PDF format or a tiff format (step S12). If it is determined that the file is an image file, since it is equivalent to the image information input in step S1, vectorization processing is performed (step S13). This vectorization process will be described later. After the vectorization process, the data is converted into data that can be edited by a general application (step S14).

  After conversion to editable data, the storage location of the electronic file is added to the electronic file as new pointer information (step S15) and stored in a predetermined storage area (step S16). Similarly, when it is determined that the image information is not image information in step S12, the storage location of the electronic file is added to the electronic file as new pointer information in step S15, and stored in a predetermined storage area in step S16. At this time, in step S15, the storage location of the electronic file is added to the electronic file stored as pointer information. As a result, the stored document becomes a search target from the next time, pointer information can be printed with a two-dimensional barcode or the like when printing an electronic file, and the stored document can be easily accessed from the printed document. . Then, this process is complete | finished.

  In the vectorization processing in step S13, conversion processing from image data to vector data is performed. Specifically, the vectorization process is performed on the object divided for each predetermined block in step S2. This vector process includes the following processes.

  -Recognize the character size, style, and font for the character block converted to code information by OCR in step S3, and make the font data visually and faithfully close to the character obtained by scanning the document. Convert.

  A character object that has not been processed in step S3 is cut out in units of one character and converted into code information by OCR.

  -The outline of a character object that is unrecognizable by OCR is traced (traced), and the outline information (outline information) is converted into a format that represents the connection of line segments.

  In the case of a graphic object, the outline information is tracked (traced) and converted into a format that is expressed as a connection of line segments.

  -The outline information in the above line segment format is converted into function information fitted by a function such as a Bezier function.

  -Recognize the shape of the figure from the outline information of the figure object, and convert it into figure definition information such as a circle, rectangle, or polygon.

  In the case of a tabular object, it recognizes ruled lines and frame lines and converts them into form format information in a predetermined format.

  In addition to these processes, various vectorization processes such as replacing raster information with predetermined command or code information may be used. These vectorization processes are performed for each object, and when the layout information of each object is stored, as described above, after being converted into data that can be edited by a general application in step S14, In S16, it is stored in the storage unit 211 as an electronic file.

  Here, the data format of an editable application is shown. Usually, the data format depends on the application to be used. Commonly known data formats include Word (registered trademark), a word processor of Microsoft Corp., and Excel (registered trademark), a spreadsheet application. There is. Each of these applications has a different purpose of use, and a file format corresponding to the purpose is defined, and a file (data) is stored in the file format.

  In addition, as a file format that can be used to some extent, there is a Microsoft Text Corp. RTF (Rich Text Format) format, in recent years an SVG (Scalable Vector Graphics) format, or a plain text format that simply handles only text data. As long as it is a corresponding application, these are file formats that can be used in common.

  In the present embodiment, it is possible to select in what file format the vectorized information is generated, and it is possible to perform an optimal vectorization process according to the selected file format. In addition, this vectorization processing not only enables editing as vector data, but also reduces the amount of information, improves the storage efficiency, and shortens the transmission time, compared to a case where image data is directly handled. Moreover, when recording and displaying, it can be handled as high-quality data.

[Block selection (BS) processing]
FIG. 5 is a diagram showing an example in which the image data stored in the storage unit 211 by the block selection process is divided into a character / line image part and a halftone image part. FIG. 4A shows image data before the block selection process is performed, and FIG. 4B shows an area separated by the block selection process. In this block selection (selection) process, the read image data for one page (see FIG. 1A) is recognized as a block for each object, and is displayed as text (TEXT), photograph (Photo), line (Line). This is a process of determining the attribute of each block, such as a table, and dividing it into areas having different attributes.

  Specifically, first, the input image is binarized into a black and white image, contour tracing is performed, and a block of pixels surrounded by a black pixel contour is extracted. For a black pixel block with a large area, the white pixel block is extracted by tracing the contours of the white pixel inside, and then the black pixel is recursively extracted from the inside of the white pixel block having a certain area or more. Extract the lump.

  The black pixel blocks obtained in this way are classified by size and shape, and are classified into areas having different attributes. For example, an area having an aspect ratio close to a value 1 and having a constant size is defined as a block of pixels corresponding to a character, and a portion where adjacent characters can be grouped in a well-aligned manner is defined as a character region. Further, a flat pixel block is defined as a line region. Further, a range occupied by a block of black pixels that are not less than a certain size and includes a block of square white pixels in a well-aligned manner is defined as a surface area. Further, an area where irregularly shaped clusters of pixels are scattered is defined as a photographic area. A block of pixels having any other shape is used as a drawing area.

  FIG. 6 is a diagram showing block information of each block obtained by the block selection process. Each block information is used as information for vectorization or search. Here, for each block, the attribute, coordinate X, coordinate Y, width W, height H, and presence / absence of OCR information are used as block information. The attribute values 1, 2, 3, 4, and 5 correspond to characters, figures, tables, lines, and photographs, respectively.

[Detection of pointer information]
Next, OCR / OMR processing for extracting the storage location of the electronic file from the image information in step S3 will be described. FIG. 7 is a flowchart showing a processing procedure for decoding a two-dimensional barcode symbol (QR code) added to a document image and outputting a data character string. FIG. 8 is a diagram showing a document to which a two-dimensional barcode is added.

  First, the image of the original 310 stored in the page memory in the data processing unit 215 is scanned by the CPU, and the position of the two-dimensional barcode symbol 311 is detected based on the result of the block selection process described above (step S21). ). Here, the QR code position detection pattern is composed of the same position detection element patterns 311 a arranged at three corners of the four corners of the symbol 311.

  The format information adjacent to the QR code position detection pattern is restored, and the error correction level and mask pattern applied to the symbol 311 are acquired (step S22). The model number of the symbol 311 is determined (step S23), and the mask process is canceled by performing an XOR operation on the encoded area bit pattern using the mask pattern obtained from the format information (step S24). In accordance with the arrangement rule corresponding to the model, the symbol character is read to restore the message data and the error correction code word (step S25).

  It is detected whether or not there is an error in the restored code (step S26). If an error is detected, it is corrected (step S27). On the other hand, if no error is detected, the process proceeds to step S28. Based on the mode indicator and the character number indicator, the data code word is divided into segments from the error-corrected data (step S28). Finally, the data character string is decoded based on the specification mode, and the result is output (step S29). Then, this process is complete | finished.

  The data incorporated in the two-dimensional barcode represents the address information of the corresponding electronic file, and is composed of path information including a file server name and a file name, for example. Alternatively, it consists of a URL to the corresponding electronic file.

  Further, in the present embodiment, the document 310 provided with pointer information using a two-dimensional barcode is shown. However, when pointer information is directly recorded as a character string, a block of character strings in accordance with a predetermined rule is displayed. It is also possible to directly obtain the address information of the original electronic file by detecting each character of the character string indicating the pointer information detected by the previous block selection process.

  In addition, pointer information can be given to the character block or character string of the original 310 by adding modulation that is difficult to visually recognize to the interval between adjacent characters and embedding information in the character interval. When character recognition processing, which will be described later, is performed based on this watermark information, pointer information is obtained by detecting the interval between characters. It is also possible to add pointer information as a digital watermark in a natural image.

[File search by pointer information]
FIG. 9 is a flowchart showing a processing procedure for searching for an electronic file based on the pointer information in step S6A. First, the file server is specified based on the address included in the pointer information (step S31). Here, the client PC 102, the document management server 106 that manages the database 105, or the multifunction peripheral 100 having the storage unit 211 is specified as the file server. The address is URL or path information including a server name and a file name.

  When the file server is specified, the address is transferred to the file server (step S32). When the file server receives the address, it searches for the corresponding file. Then, the search result is received from the file server (step S33). It is determined whether or not there is a file (step S34). If the file does not exist, this is notified to the multi-function device 100, the process is terminated, and the process returns to the main process. On the other hand, if the file exists, the file address is notified and the file server is instructed to transfer the file (step S35). Thereafter, the present process is terminated and the process returns to the main process.

  In the main process shown in FIGS. 3 and 4, based on the search result of step S33, as described above, it is determined whether or not the corresponding electronic file exists in step S6, and the corresponding electronic file does not exist. , To that effect, the process proceeds to step S7. On the other hand, if the corresponding electronic file exists, the electronic file address is notified in step S10, and the electronic file transferred from the file server is received in step S11.

[File search processing]
10 and 11 are flowcharts showing the file search processing procedure in step S7. As described above, the processing in step S7 is performed when the pointer information does not exist in the image information (input file) of the document input in step S5, or even if the pointer information exists in step S6. It is executed when not found.

  Here, a case is shown where each block and input file extracted as a result of the processing in step S3 have the information (block information and input file information) shown in FIG. As described above, in FIG. 6, the contents of the block information include attributes, coordinate positions, width and height sizes, and presence / absence of OCR information.

  Attributes are classified into characters, lines, photographs, pictures, tables and others. Also, here, in order to simplify the description, each block has a block 1, a block 2, a block 3, a block 4, a block, in order of increasing coordinates X (X1 <X2 <X3 <X4 <X5 <X6). 5 and block 6 are numbered. The total number N of blocks is the total number of blocks in the input file. Here, the total number N of blocks is the value 6.

  Then, using these pieces of information (block information and input file information), a layout search for a file similar to the input file is performed from the database. In the process of FIG. 10, the input file and the file in the database are sequentially compared. First, initialization such as a similarity ratio is performed on a file in the first database (step S41). Then, the total number of blocks is compared (step S42). The comparison of the number of blocks is performed according to the mathematical formula (1). Here, n is the total number of blocks in the file in the database. ΔN is an error in the total number of blocks in the input file.

N−ΔN <n <N + ΔN (1)
If the formula (1) is satisfied in step S42, the block information in the file is further compared sequentially. In this block information comparison, an attribute similarity rate, a size similarity rate, and an OCR similarity rate are calculated, and an overall similarity rate is calculated based on these values. That is, the block attributes of the input file and the file in the database are compared (step S43), and if they match, the attribute similarity of the file in the database is updated (step S44). Further, it is determined whether or not the sizes of the input file and the file in the database satisfy the formula (2), that is, a size comparison process is performed (step S45). Here, W is the block width of the input file. H is the block height of the input file. ΔW is an error in the block width of the input file. ΔH is the block height of the input file. w is the block width of the file in the database. h is the block height of the file in the database.

W−ΔW <w <W + ΔW and H−ΔH <h <H + ΔH (2)
If the mathematical formula (2) is satisfied, the file size similarity in the database is updated (step S46). Further, it is determined whether or not both the input file and the file in the database have OCR information (step S47). If it has OCR information, the OCR information is compared (step S48), and the OCR similarity is updated (step S49). In addition, since a well-known technique is used for calculation of each similarity rate of an attribute, size, and OCR information, the description is abbreviate | omitted.

  And it is discriminate | determined whether the similarity rate comparison was performed about all the blocks of an input file (step S50). If all the blocks are not finished, the process returns to the process of step S43 after performing the process of step S55. In step S55, if N ≦ n, that is, if the total block number n of the file in the database is greater than or equal to the total block number N of the input file, the next block in the input file is processed, while N> n. In some cases, that is, when the total block number n of the file in the database is smaller than the total block number N of the input file, the next block of the file in the database is processed. If the input file and the block attribute of the file in the database do not match in step S43, the size of the input file and the file in the database does not satisfy Expression (2) in step S45, and the input file in step S47. If the file in the database does not have OCR information, the process of step S55 is similarly performed, and then the process returns to step S43.

  When the comparison of the similarity ratios is completed for all the blocks of the input file in step S50, the total similarity ratio is calculated (step S51). It is determined whether or not the calculated total similarity is higher than a preset threshold Th (step S52). If the calculated total similarity is higher than a preset threshold Th, the file is stored as a similarity candidate (step S53). Thereafter, it is determined whether or not all files in the database have been completed (step S54). If all the files in the database have not been terminated, the next file is set as the processing target (step S56), and the process returns to step S41. If the total similarity is equal to or smaller than the preset threshold value Th in step S52, the next file is set as a processing target (step S56), and the process returns to step S41. On the other hand, when all the files in the database are finished in step S54, this process is finished. In the size comparison process in step S45, the position information may be compared based on the coordinates X and Y.

  As a result of the search, the total similarity (similarity) is higher than the threshold value Th, and the files in the database stored as similar candidates in step S53 are displayed as thumbnail images in step S8 described above. Then, a file is specified from among a plurality of similar candidates as required by an operator input operation.

[Vectorization processing]
Next, vectorization processing in step S13 will be described. In this process, when the electronic file cannot be specified in step S9, the entire image is vectorized. First, character recognition processing is performed for each character in the character block.

(Character recognition)
In the character recognition processing, recognition processing is performed on an image cut out in character units using a pattern matching method, and a corresponding character code is acquired. This recognition process compares an observed feature vector obtained by converting a feature obtained from a character image into a numerical sequence of several tens of dimensions with a dictionary feature vector obtained in advance for each character type, and determines the character type with the closest distance. This is a process for obtaining a recognition result. Various methods are known for extracting feature vectors. For example, a method that features a mesh number-dimensional vector that divides characters into meshes and counts the character lines in each mesh as line elements by direction. It has been known.

  When character recognition is performed on the character area extracted by the block selection process in step S2, first, horizontal / vertical writing is determined for the corresponding area, and a line is cut out in each corresponding direction. To obtain a character image. Here, the horizontal / vertical writing is determined by taking a horizontal / vertical projection of the pixel value in the corresponding area, and determining that the horizontal projection area is large when the horizontal projection dispersion is large. This is done by determining that it is a writing area. In the case of horizontal writing, the character string and the character are decomposed by cutting out a line using a horizontal projection, and further cutting out a character from a vertical projection with respect to the cut out line. On the other hand, the character string and the character for the vertically written character area may be decomposed into characters and characters in the same manner as in the horizontal writing with the horizontal and vertical reversed. At this time, the character size is detected.

(Font recognition)
By preparing multiple dictionary feature vectors for the number of character types used for character recognition for the character shape type (font type), and outputting the font type together with the character code at the time of matching, the font of the character Recognizable.

(Character vectorization)
Based on the character code and font information obtained by character recognition and font recognition, the character portion information is converted into vector data using outline data prepared in advance. When the original document is in color, the color of each character is extracted from the color image and recorded together with vector data. Thereby, the image information belonging to the character block can be converted into vector data in which the shape, size and color are reproduced almost faithfully.

(Vectorization of parts other than characters)
In the block selection process, a region separated as a figure, an image, a line, and a table region is targeted, and the contour of the pixel block extracted therein is converted into vector data. Specifically, a point sequence of pixels forming an outline is divided by points regarded as corners, and each section is approximated by a partial straight line or curve. FIG. 12 is a diagram illustrating a state in which corners are approximated by curves. Here, the corner is a point where the curvature becomes a maximum. The point where the curvature is maximum is obtained as the point where the distance between the string and the point Pi becomes maximum when a string is drawn between the left and right points Pi-k and Pi + k with respect to the arbitrary point Pi. It is done. Furthermore, let R be the chord length / arc length between the points Pi-k and Pi + k, and the point where the value of R is equal to or less than the threshold value can be regarded as a corner. In each section after being divided by the corners, a straight line is vectorized using a least square method for a point sequence, and a curve is vectorized using a cubic spline function or the like. When the target has an inner contour, the inner contour is similarly partially approximated by a straight line or a curve using the point sequence of the white pixel contour extracted by the block selection process.

  In this way, the outline of a figure having an arbitrary shape can be vectorized by using the contour lane marking approximation. When the original document is in color, the color of the figure is extracted from the color image and recorded together with the vector data.

  FIG. 13 is a diagram illustrating a state of expressing as a line having a thickness. When an outer contour and an inner contour or another outer contour are close to each other in a certain section, two contour lines can be combined and expressed as a line having a thickness. Specifically, when a line is drawn from each point Pi of a certain contour to a point Qi that is the shortest distance on another contour, and each distance PQi is equal to or less than a certain length on average, the section of interest is defined as a line segment PQi. Is approximated by a straight line or a curve having a middle point as a point sequence, and its thickness is taken as the average value of the line segments PQi. A table ruled line, which is a line or a set of lines, can be efficiently expressed as a set of lines having such a thickness by a vector.

  As described above, when vectorization is performed using the character recognition process in the character block, the character having the closest distance from the dictionary as a result of the character recognition process is used as the recognition result. When this distance is greater than or equal to a predetermined value, it is not always the same as the original character, and is often erroneously recognized as a character having a similar shape. Therefore, in the present embodiment, such a character that is easily misrecognized is handled in the same manner as a general line drawing, and the character is outlined. Thereby, conventionally, even a character that is likely to be erroneously recognized in the character recognition process is not vectorized into an erroneous character, and can be vectorized by an outline that is visually faithful to image data. In addition, a block determined to be a photograph cannot be vectorized, and thus remains as image data.

(Figure recognition)
Next, processing for vectorizing the outline (outline) of a figure having an arbitrary shape and then grouping the vectorized dividing lines for each graphic object will be described. FIG. 14 is a flowchart showing a processing procedure for grouping vector data for each graphic object. First, the start point and end point of each vector data are calculated (step S61). A graphic element is detected using the calculated start point and end point information of each vector data (step S62). Here, the detection of the graphic element means detecting a closed graphic formed by the dividing line. When detecting this closed figure, the principle that each vector which comprises a closed shape (closed figure) has the vector respectively connected with the both ends is applied. Then, other graphic elements or dividing lines existing in the graphic element are grouped into one graphic object. On the other hand, if there is no dividing line between other graphic elements in the graphic element, the graphic element is regarded as a graphic object. (Step S63). Then, this process is complete | finished.

  FIG. 15 is a flowchart showing a processing procedure for detecting a graphic element. First, based on the vector data, unnecessary vectors not connected to both ends are removed, and a closed graphic component vector is extracted (step S71). From the closed figure constituent vector, the starting point of this vector is set as the starting point, and the vectors are sequentially followed clockwise. Tracking is performed until the start point is returned again, and all the passed vectors are grouped as closed figures constituting one figure element (step S72). In the group processing in step S72, all the closed graphic constituent vectors inside the closed graphic are also grouped. Further, the same processing is repeated with the starting point of the vectors not yet grouped as the starting point. Thereafter, among the unnecessary vectors removed in step S71, those joined to the vector grouped as a closed figure in step S72 are detected and grouped as one figure element (step S73). Then, this process is complete | finished. Thereby, it is possible to handle the graphic block as an individual graphic object that can be reused individually.

(Conversion to application data)
In the block selection process (see step S2), the image data for one page is blocked and vectorized (see step S13). As a result, the image data is converted into an intermediate data format file. This intermediate data format is referred to as a document analysis output format (DAOF) format. FIG. 16 is a diagram showing a data structure in the DAOF format. The data structure of the DAOF format includes a header 791, a layout description data portion 792, a character recognition description data portion 793, a table description data portion 794, and an image description data portion 795.

  A header 791 holds information regarding document image data to be processed. The layout description data portion 792 includes attributes such as TEXT (character), TITLE (title), CAPTION (caption), LINEART (line drawing), PICTURE (natural image), FRAME (frame), and TABLE (table) in the document image data. The attribute information of each block recognized every time and its rectangular address information are held. The character recognition description data portion 793 holds a character recognition result obtained by character recognition of a TEXT block such as TEXT, TITLE, and CAPTION. The table description data portion 794 stores details of the structure of the TABLE block. The image description data portion 795 cuts out image data of blocks such as PICTURE and LINEART from the document image data and holds them.

  Such a DAOF is not only used as intermediate data but may be stored as a file itself. In this file state, it is not possible to reuse individual objects by a so-called general document creation application. Can not. Therefore, a process (see step S14) for converting the data in the DAOF format into application data (simply referred to as application data) will be described.

  FIG. 17 is a flowchart showing the application data generation processing procedure in step S14. First, DAOF is input (step S81). A document structure tree that is the source of application data is generated (step S82). Based on the document structure tree, actual data is inserted into the DAOF to generate actual application data (step S83). Then, this process is complete | finished.

  FIG. 18 is a flowchart showing a document structure tree generation processing procedure. FIG. 19 shows a document structure tree. As a basic rule of overall control, processing is performed so as to shift from a micro block (a single block) to a macro block (an aggregate of blocks). Here, the block refers to the micro block and the entire macro block.

  First, regrouping is performed on a block-by-block basis based on the vertical relationship (step S91). Immediately after the start, a determination is made in units of micro blocks. Here, the relevance is defined by the fact that the distance is close and the block width (the block height in the case of the horizontal direction) is substantially the same. Information such as distance, width, and height is extracted with reference to DAOF.

  FIG. 19A shows the actual page configuration, and FIG. 19B shows the document structure tree. As a result of the grouping in step S91, the block V1 in which the blocks T3, T4, and T5 form one group and the block V2 in which the blocks T6 and T7 form one group are generated as a group in the same hierarchy.

  The presence / absence of a vertical separator is checked (step S92). Here, the separator is, for example, an object physically having a line attribute in the DAOF. Also, logically, it is an element that explicitly divides a block in an application. If a separator is detected, it is subdivided at the same level.

  Using the group length in the vertical direction, it is determined whether or not the grouping length in the vertical direction is the page height, that is, whether there are no more divisions (step S93). When the group length in the vertical direction is the page height, the document structure tree generation process ends.

  On the other hand, if the group length in the vertical direction is not the page height, the process proceeds to step S94. In FIG. 19, since there is no separator and the group height is not the page height, the process proceeds to step S94.

  Then, regrouping is performed on a block basis based on the relevance in the horizontal direction (step S94). Also in this case, in the first time immediately after the start, the determination is performed in units of micro blocks. The definition of the relevance and the determination information is the same as in the vertical direction. In FIG. 19, blocks T1 and T2 are generated as H1 groups in the same hierarchy one above blocks T1 and T2, and blocks V1 and V2 are generated as H2 groups in the same hierarchy one above blocks V1 and V2.

  The presence / absence of a horizontal separator is checked (step S95). In FIG. 19, since the separator S1 exists, it is registered in the tree, and a hierarchy of blocks H1, S1, and H2 is generated.

  Using the group length in the horizontal direction, it is determined whether or not the grouping length in the horizontal direction is the page width, that is, whether there are no more divisions (step S96). If the horizontal group length is the page width, the document structure tree generation process is terminated.

  On the other hand, if the group length in the horizontal direction is not the page width, the process returns to step S91, and the same process is repeated from the relevance check in the vertical direction for the next higher layer again. In FIG. 19, since the division width is the page width, the processing ends here, and finally the highest-level block V0 representing the entire page is added to the document structure tree.

  After the document structure tree is completed, application data is generated in step S83 based on the information. In the case of FIG. 19, since the block H1 has two blocks T1 and T2 in the horizontal direction, it is output as two columns, and internal information of the block T1 (DAOF reference, text of character recognition result, image, etc.) is output. After that, the column is changed, the internal information of the block T2 is output, and then the separator S1 is output. In addition, since there are two blocks V1 and V2 in the horizontal direction in the block H2, the data is output as two columns. In the block V1, the internal information is output in the order of the blocks T3, T4, and T5, and then the columns are changed. The internal information of the blocks T6 and T7 of the block V2 is output. In this way, conversion processing to application data is performed.

[Add pointer information]
Next, pointer information addition processing in step S15 will be described. When the document to be processed is specified by the search process, or when the original file can be reproduced by vectorization, this printing is performed by adding pointer information when recording the document (recording on paper). When performing various processes again using a document, the original file data can be easily obtained.

  FIG. 20 is a flowchart showing the pointer information addition processing procedure in step S15. In this pointer information addition processing, a data character string as pointer information is encoded into a two-dimensional barcode symbol (QR code: JIS X0510) 311 and added to the image. The data incorporated in the two-dimensional barcode represents the address information of the corresponding electronic file, and is composed of, for example, path information including a file server name and a file name. Alternatively, the URL includes a URL to the corresponding electronic file, the database 105 or 115 in which the corresponding electronic file is stored, or a file ID managed by the storage unit 211 in the multifunction peripheral 100.

  First, in order to identify various different characters to be encoded, the input data string is analyzed (step S101). Further, an error detection and error correction level is selected, and a minimum model number that can accommodate input data is selected.

  An input data string is converted into a predetermined bit string, and an indicator or termination pattern indicating a data mode (numeric, alphanumeric, 8-bit byte, kanji, etc.) is added as necessary. Further, it is converted into a predetermined bit code word (step S102). In order to perform error correction, the code word string is divided into a predetermined number of blocks according to the model number and the error correction level, an error correction code word is generated for each block, and is added after the data code word string (step S103). .

  The data code words of the respective blocks obtained in step S103 are connected, and the error correction code word of each block and, if necessary, the remainder code word are followed (step S104). A code word module is arranged in the matrix together with a position detection pattern, a separation pattern, a timing pattern, an alignment pattern, and the like (step S105).

  An optimal mask pattern is selected for the symbol coding region, and the mask processing pattern is converted into a module obtained in step S105 by an XOR operation (step S106). Format information and model number information are generated in the module obtained in step S106 to complete a two-dimensional code symbol (step S107). Then, this process is complete | finished.

  The two-dimensional barcode in which the address information is incorporated in this way is raster data that can be recorded by the data processing unit 215 when the client PC 102 forms an electronic file as print data and as a recorded image on a sheet of the recording unit 212, for example. Then, the image is added to a predetermined location on the raster data to form an image. Here, the user to whom the paper on which the image is formed is distributed can detect the storage location of the original electronic file from the pointer information in step S4 by reading the image with the image reading unit 210.

  In addition to the two-dimensional bar code shown in the present embodiment, for example, a method of adding pointer information directly to a document as a character string, a character string in a document, particularly a character and character In general, a method called digital watermark can be applied, such as a method of embedding information by modulating an interval, or a method of embedding in a halftone image in a document.

[Select and specify file format]
Even if the original electronic file cannot be specified by the above processing, an electronic file equivalent to that created by the original application can be obtained. However, the file format of the vectorized electronic file generated here is limited to a predetermined format. There are a wide variety of applications that are actually assumed, and the convenience of the user will vary greatly depending on the type of computerization.

  Therefore, in this embodiment, a user interface is provided so that the user can select a file format to be generated. This selection process by the user is performed before the start of the vectorization process in step S13 or before the start of the image information input process in step S1. FIG. 21 is a diagram showing an operation screen for selecting an output file format. The operation screen of FIG. 21 shows a message displayed on the user operation screen of the multifunction device 100 or the management PC 101. Here, an “output file format selection screen” is displayed, and the user can specify in which format the electronic data subjected to vectorization processing is to be written. It is possible to select from a plurality of file formats, and an operation is performed so as to write in a specified format by an ON or OFF check box. Here, a “plain text format”, “word processor format”, “table format”, “presentation format”, “SVG”, or “RTF” format can be selected. In this embodiment, the WORD (registered trademark) format is exemplified as the word processor format, the Excel (registered trademark) format is exemplified as the table format, and the PowerPoint (registered trademark) format is exemplified as the presentation format, but other formats may be used. It is. When a plurality of formats are specified, a plurality of electronic data according to each format is output.

  Here, since the WORD (registered trademark) format and the SVG format are selected, the generated vectorized data is output to the data storage area in these two formats. In the present embodiment, an optimal vectorization process is applied to each selected format.

  For example, when only the plain text format is selected, the only data that can be handled in this format is character code information. Therefore, the vectorization processing is applied only to the minimum level, and only the character image encoding processing by OCR is performed. That is enough.

  In addition, when the WORD (registered trademark) format or the RTF format is selected, character code information, tabular frame / ruled line information, and bitmap image information can be mixedly handled. Table data frame and ruled line vectorization, layout bitmap image cutout processing, and the like are executed. On the other hand, the function of the figure outline is not approximated.

  When the SVG format is selected, almost all vectorization information can be defined. Therefore, all the vectorization processes described above are executed, and the vectorized data is converted into the SVG language and output. To do.

  In addition to controlling execution / non-execution of processing, it is possible to assume a case where the priority of processing is changed. For example, when the Excel (registered trademark) format is selected, it is expected that the input image is often a table format expressed by vertical ruled lines and horizontal ruled lines. For each object, priority is given to the process of extracting vertical and horizontal straight lines. As a result of processing in this manner, even in graphic parts other than the table, vertical / horizontal straight line parts are easily extracted and the extraction accuracy as a block of figures is slightly lowered, but the extraction accuracy of the table is increased. As a result, a preferable output can be obtained.

  In addition, when the PowerPoint (registered trademark) format is selected, it is expected that a lot of graphic information such as rectangles and circles is included. Therefore, it is possible to preferentially perform such graphic extraction processing. is there. In this case, graphic information such as circles and rectangles is output as command format data (Circle, Square, etc.) that defines them.

FIG. 22 is a diagram showing a changing method of vectorization processing. In this example, a document image including a plurality of objects numbered by solid lines is shown on a page 801. The entire page 801 is input as digital raster image data from the image reading unit 210 and divided into objects by BS processing (see step S2), and then each object becomes a vectorization processing target. It shows what kind of vectorization processing each of these objects is subjected to depending on the output file format.
(A) When plain text format is specified In this case, only a character string is subject to vectorization. In the figure, only the objects of the respective character strings indicated by reference numerals 803, 806, 809, and 813 are subject to vectorization, and are character-coded by OCR processing. Other objects are ignored.
(B) When PowerPoint, WORD, or RTF format is specified Character encoding and replacement with basic graphics are performed. The character string is converted into a character code by OCR processing as in (a) above. The elliptical objects 802 and 812 are replaced with an elliptical figure having an optimum major axis / minor axis based on the outline information of the figure. The rectangular object 805 and the diamond object 808 are similarly converted. The arrows 804, 807, and 811 are defined as straight lines having an arrow at the end point because they are replaced with straight line vectors based on the coordinates of the start point and end point, and an arrow is added at the end. The curve 810 is vectorized as a set of short straight lines based on the outline from the start point to the end point, and similarly, an arrow mark is given to the end point.

(C) When SVG format is selected The character string processing is the same as (a) and (b) above, but here, font shape vectorization processing is also performed as processing unique to SVG. That is, the character string is decomposed into character units, and the SVG font definition is generated based on the outline information of each character. This font information is defined integrally with the character code information by OCR. By performing such processing, even a document in which a special font format is used for a character string can be vectorized while the character shape is preserved.

  Further, it is possible to perform Bezier function approximation of outline data for each graphic object 802, 805, 808, and 812 without converting to graphic data. Similarly, a smooth curve can be vectorized as it is by approximating the outline to the curve 810 with a Bezier function. The processing of the arrow is the same as (b) above.

  As described above, various vectorization processes can be appropriately performed according to the designated output file format. In addition, the reusability of the generated vector data by using the application is further enhanced. The above processing method is merely an example, and various other processing methods can of course be considered. Each selected file format has its own document structure tree information, header information, and data storage format. Of course, application data in a format corresponding to the selected format is generated. is there.

  According to the image processing system of the first embodiment, when specifying an electronic file of an original from image information obtained by reading and scanning the original, if the electronic file cannot be specified, a vector is calculated from the image information. Data can be generated and converted to application files. Therefore, it is possible to obtain the same effect as if the original electronic file was obtained. In addition, an application file desired by the user can be generated by designating the format of the application file by the user. In addition, the vectorization process performed at this time is set so as to be optimal for the designated application, so that a conversion result desirable for the user can be obtained.

[Second Embodiment]
Some document files to be handled should limit reuse by third parties. In the first embodiment, it is assumed that all the files stored in the file server can be freely accessed, and the entire file or a part of its objects can be reused. Therefore, in the second embodiment, when a file is searched from pointer information, a case is shown in which the access right is restricted for the specified file as a result of the search. FIG. 23 is a flowchart showing a processing procedure for searching for an electronic file based on the pointer information in step S6A of the second embodiment. Since the processing from step S121 to S124 is the same as the processing from step S31 to S34 in FIG. 9 of the first embodiment, the description thereof is omitted.

  If the file is specified in step S124, the file server is checked for the presence of access right information for the file (step S125). If there is an access restriction, upon receiving a password transmission request (step S126), the multifunction device prompts the operator to input the password, and transmits the input password to the file server (step S127).

  The verification result of the transmitted password by the file server is determined (step S128). As a result of the verification, if the password does not match and the authentication fails, the process returns to step S126. On the other hand, if the password matches and the authentication is successful, the file address is notified, and if the user-desired process is acquisition of image file data, the file server is instructed to transfer the corresponding file (step S129). . Thereafter, this process is terminated and the process returns to the main process.

  The authentication method for controlling the access right is not limited to the password method shown in steps S126 and S127. For example, all authentication methods such as fingerprint authentication and widely used biometric authentication and card authentication are available. Authentication means can be used. In the second embodiment, the embodiment is shown in which the file is specified by the pointer information given to the paper document. However, as shown in steps S7 and S8, the file is specified by the file search process. The same applies to the above.

  On the other hand, when the file cannot be specified by the file server, it is possible to add a restriction to the vectorization process in step S13. In other words, when the presence of a restriction on the access right to this paper document is detected from the image information obtained by scanning the paper document, the vectorization process is performed only when authentication is confirmed, thereby providing high confidentiality. Limit the use of documents.

[Third Embodiment]
In the first embodiment, when the original electronic file cannot be specified by the file search process, the vectorization process is performed on the entire image image. For example, in the case of a general document, all objects in the document are processed. Are not newly created, and some objects may be created by diverting from other files.

  For example, as a background object (wallpaper), several patterns are prepared in advance by a document creation application, and it is normal to select and use them. Such an object is likely to exist in other document files in the document file database, and is likely to exist as reusable vector data.

  Therefore, from such a background, as a vectorization process in step S13, for each object divided into individual objects in the block selection process, a file partially including a matching object from the database in this object unit Are retrieved, and vector data is obtained for each matching object individually from the file in units of objects. As a result, it is not necessary to vectorize the entire document, vectorization processing can be performed at a higher speed, and image quality deterioration due to vectorization can be prevented.

  Further, in the third embodiment, when the output file format is specified as shown in FIG. 21, it is possible to limit the search processing target to the selected format file. For example, when the PowerPoint (registered trademark) format is selected, the processing is simplified by searching and acquiring the vector data in units of objects for only the PowerPoint (registered trademark) file on the file server. be able to.

  On the other hand, in the file search process in step S7, when the original file is specified as a PDF file, this PDF file has a character code that has already been recognized for the character object of the document as an additional file. There is a case. When vectorizing such a PDF file, the character recognition process in the vectorization process after step S13 can be omitted by using this character code file. That is, the vectorization process can be performed at a higher speed.

  The above is the description of the embodiments of the present invention. However, the present invention is not limited to the configurations of these embodiments, and the functions shown in the claims or the functions of the configurations of the embodiments are included. Any configuration that can be achieved is applicable.

  Another object of the present invention is to supply a storage medium storing software program codes for realizing the functions of the embodiments to a system or apparatus, and a computer (or CPU, MPU, etc.) of the system or apparatus stores the storage medium. It is also achieved by reading out and executing the program code stored in. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

  Examples of storage media for supplying the program code include ROM, floppy (registered trademark) disk, memory card such as PCMCIA card and compact flash (registered trademark), hard disk, micro DAT, magneto-optical disk, CD- It may be composed of an optical disk such as R or CD-RW, a phase change optical disk such as DVD, or the like.

  Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but also an OS (operating system) running on the computer based on the instruction of the program code. A case where part or all of the actual processing is performed and the functions of the above-described embodiments are realized by the processing is also included.

  Further, after the program code read from the storage medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, the function expansion is performed based on the instruction of the program code. This includes the case where the CPU or the like provided in the board or the function expansion unit performs part or all of the actual processing and the functions of the above-described embodiments are realized by the processing.

It is a figure which shows the structure of the image processing system in 1st Embodiment. 2 is a diagram illustrating an electrical configuration of the multifunction peripheral 100. FIG. 3 is a flowchart illustrating an operation processing procedure of the multifunction peripheral 100. FIG. 4 is a flowchart illustrating an operation processing procedure of the multifunction peripheral 100 continued from FIG. It is a figure which shows the example which isolate | separates the area | region of the image data preserve | saved at the memory | storage part 211 by the block selection process into the character / line drawing part and the halftone image part. It is a figure which shows the block information of each block obtained by the block selection process. It is a flowchart which shows the process sequence which decodes the two-dimensional barcode symbol (QR code) added in the original image, and outputs a data character string. It is a figure which shows the original document to which the two-dimensional barcode was added. It is a flowchart which shows the process sequence which searches an electronic file based on the pointer information in step S6A. It is a flowchart which shows the file search processing procedure in step S7. It is a flowchart which shows the file search process procedure in step S7 following FIG. It is a figure which shows a mode that an angle is approximated with a curve. It is a figure which shows a mode that it expresses as a line with thickness. It is a flowchart which shows the process sequence which groups vector data for every graphic object. It is a flowchart which shows the process sequence which detects a graphic element. It is a figure which shows the data structure of DAOF format. It is a flowchart which shows the production | generation process procedure of the application data in step S14. It is a flowchart which shows a document structure tree production | generation process procedure. It is a figure which shows a document structure tree. It is a flowchart which shows the addition process procedure of the pointer information in step S15. It is a figure which shows the operation screen which selects an output file format. It is a figure which shows the change system of a vectorization process. It is a flowchart which shows the process sequence which searches an electronic file based on the pointer information in step S6A of 2nd Embodiment.

Explanation of symbols

100 MFP 101 Management PC
102 Client PC
105, 115 Database 106, 116 Document management server 210 Image reading unit 211 Storage unit 212 Recording unit 213 Input unit 215 Data processing unit 216 Display unit 311 Two-dimensional barcode symbol (QR code)

Claims (10)

Vectorization means for generating vector data from image data input from input means for inputting image data;
An image processing apparatus comprising: output means for outputting the generated vector data in a predetermined file format. Selecting means for selecting the predetermined file format;
The image processing apparatus according to claim 1, wherein the vectorization unit changes content of vectorization processing when generating the vector data based on the selection result.   The image processing apparatus according to claim 1, wherein the input unit optically reads an image of a document to generate an image signal, and converts the generated image signal into image data. Read means described in the converted image data and reading pointer information indicating a storage location of the original electronic file,
4. The image according to claim 3, wherein when the electronic file acquired based on the read pointer information is image data, the vectorization means generates the vector data from the converted image data. Processing equipment.   The image processing apparatus according to claim 1, wherein the input unit inputs an array of two-dimensional digital pixel data as the image data.   The image processing apparatus according to claim 1, wherein the vectorization unit converts the input raster format image data into command definition format data such as code information, graphic information, and function information.   2. The image processing apparatus according to claim 1, wherein the vectorization means includes format conversion means for converting an object of the image data converted into the vector data into a predetermined format that can be handled by predetermined application software. .   8. The image processing apparatus according to claim 7, wherein the predetermined format is a designated file format or a document structure format defined by software that generates the file format. An input step for inputting image data;
A vectorization step of generating vector data from the input image data;
And an output step of outputting the generated vector data in a predetermined file format. A program executed by a CPU in the image processing apparatus,
An input step for inputting image data;
A vectorization step of generating vector data from the input image data;
An output step of outputting the generated vector data in a predetermined file format.

Images