US 5,659,767 AGrant
Application Programming Interface for Accessing Document Analysis Functionality of a Block Selection Program
Issue Date:1997-08-19
•17 Claims
•7 Drawing Sheets
Abstract
An application programming interface accesses page analysis functionality of a block selection program and accesses data structures resulting therefrom. The application programming interface includes plural page analysis functionality which manipulate block selection functions including functions to analyze a document page, to create a hierarchical tree structure corresponding to the analyzed document page, and to return a pointer which is a block ID or a root node of the analyzed page. Plural hierarchical tree structure accessing functions traverse the hierarchical tree structure, and locate desired nodes in the hierarchical tree structure.
Metadata
Assignee
- Canon Information Systems, Inc.
Inventors
- Thieu Quang Nguyen
- Mann Chang
Application Information
Application Number:US 3387666
Filing Date:1994-11-10
Priority Date:1994-11-10
Art Unit:237
Classifications
IPC:
G06F 1730
Field of Search:
395382364500;600;776;77716;22419.19
Patent Drawings (7 sheets)
Description
Background of the Invention
This application is being filed with an appendix of computer program listings.
A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objections to the facsimile reproduction by any one of the document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.
Field of the Invention
The present invention relates to an application programming interface (API) for accessing document analysis functionality of a block selection program and for accessing the data structures resulting therefrom. More particularly, the present invention relates to a page analysis API library and a method for making a page analysis program utilizing the API library, whereby the functions in the API library can traverse and access data in a hierarchical tree structure resulting from a block selection technique.
Incorporation by Reference
U.S. applications Ser. No. 07/873,012, "Method And Apparatus For Character Recognition" Ser. No. 08/171,720, now U.S. Pat. No. 5,588,072, "Method And Apparatus For Selecting Text And/Or Non-Text Blocks In A Stored Document", and Ser. No. 08/338,781, "Page Analysis System", are incorporated herein by reference.
Description of the Related Art
Recently developed block selection techniques, such as the techniques described in U.S. patent application Ser. Nos. 07/873,012 and 08/171,720, automatically analyze images within a document page in order to distinguish between different types of image data within the document page. The result of such a block selection technique is used to determine the type of subsequent processing to be performed on the image, such as optical character recognition (OCR), data compression, data routing, etc. For example, image data which is designated as text data will be subjected to OCR processing while image data, which is designated as picture data, would not be subjected to OCR processing. As a result, the different types of image data can be input and automatically processed without an operator's intervention.
An example of how a block selection technique operates, such as the ones referenced above, will be discussed below with respect to FIGS. 1-3.
FIG. 1 shows the page of a representative document. Document page 101 is arranged in a two-column format. The page includes title 102, horizontal line 104, text areas 105, 106 and 107, which include lines of text data, halftone picture area 108, which includes a graphic image which is non-text, table 110, which includes text information, framed area 116, halftone picture area 121 accompanied by caption data 126, and picture areas 132 and 135 accompanied by caption data 137. According to the block selection techniques described in U.S. patent application Ser. Nos. 07/873,012 and 08/171,720, now U.S. Pat. No. 5,588,072 the disclosures of which are hereby incorporated by reference, each area of document page 101 is designated in accordance with a type of image data obtained therein and image data is then segmented based on its respective type. As the block selection program processes the document page, a hierarchical tree structure is created as shown in FIG. 2.
As shown in FIG. 2, hierarchical tree structure 200 contains a plurality of nodes which represent segmented blocks of image data. Each node of the tree contains feature data which defines the feature of each block of image data in the processed document page. For example, the feature data may include block location data, size data, attribute data (image type, such as text, picture, table, etc.), sub-attribute data, and child node and parent node pointers. In the present invention, child or "descendent" nodes represent image data which exist entirely within a larger block of image data. Child nodes are depicted in the hierarchical tree structure as a node branching from a parent node, such as the nodes at the same level as node 211, which branch from parent or root node 201. In addition to the feature data described above, a node which represents a text block may also contain feature data defining the block's reading orientation and reading order.
Once a hierarchical tree structure, such as hierarchical tree structure 200, has been created, it is stored in memory. Upon receiving a request to process the image data in the document image, hierarchical tree structure 200 is retrieved from memory and image data which has been blocked together is processed according to the feature data stored in its corresponding node.
In addition to using the hierarchical tree structure to process image data of a document page, the hierarchical tree structure can be used to generate and display a comprehensible format of document page 101 as shown in FIG. 3. As a result, block template 301 of the document page is generated and displayed to the user based on the feature data stored in the hierarchical tree structure shown in FIG. 2.
The block template of the document page directly reflects the structure of the hierarchical tree as well as the feature data stored in each node of the hierarchical tree structure. The feature data in the node is utilized to generate the block outline, location, and size. In addition, feature data in each node is used to identify the type of image data contained within the block and, if appropriate, reading order and reading orientation. For example, as shown in FIG. 3, block template 301 includes text blocks 302, 304, 305 and 309, each of which corresponds to nodes 202,204, 205 and 209, respectively. As shown, each text block includes feature data which designate the block as text and also define a block's reading order and reading orientation.
However, in order for the user to utilize functions of the block selection program or to access information in the resulting hierarchical tree structure, the user must be familiar with the low level process steps of the block selection program as well as how data is stored in the resulting hierarchical tree structure. That is, application developers, in order to interface with the block selection program, must understand the low level process steps of the block selection program as well as the tree structure and node information. Consequently, if the low level implementation of the block selection process steps and resulting hierarchical tree structure are changed, all applications which have been developed to interface at a low level interaction must be changed accordingly.
There is, therefore, a need to develop an API which includes functions and data files which insulate the user from the tree and node details such that the functions and data files of API may remain the same regardless of the low level implementation of the block selection technique used.
Summary of the Invention
It is an object to the present invention to provide an application programming interface (API) for page analysis to use functions of a block selection program and to access a data structure resulting therefrom.
In one aspect of the present invention, the present invention is an application programming interface for accessing page analysis functionality of a block selection program and for accessing data structures resulting therefrom. The application programming interface includes plural block selection functions which can manipulate block selection functionality that include functions to analyze a page of a document, create a hierarchical tree structure based on the page analysis, and to return a pointer which is a block ID of a root node of the hierarchical tree structure which represents the analyzed page. The application programming interface also includes plural hierarchical tree accessing functions to traverse the hierarchical tree structure and to obtain pointers to nodes within the hierarchical tree structure which contain image data attributes for corresponding blocks of image data requested by the user, and plural block functions for examining and obtaining node pointers to nodes which include a desired image data attribute.
According to another aspect of the present invention, the present invention is a method of creating an application programming interface for accessing page analysis functionality of a block selection program and for accessing a data structure resulting therefrom. The method includes storing, in memory, plural page analysis functionality which manipulate block selection functions, and plural hierarchical tree structure accessing functions. The method further includes the steps of selecting, from memory, at least one plural page analysis functionality and at least one plural hierarchical tree structure accessing functions, and combining the at least one plural page analysis functions and the at least one plural hierarchical tree structure accessing functions so as to create the application programming interface whereby the created application programming interface is capable of accessing page analysis functionality of a block selection program and is capable of accessing a data structure resulting therefrom.
This summary of the invention has been provided so that the nature of the invention may be understood quickly. A more complete understanding of the invention and its advantages may be obtained by reference to the following detailed description in connection with the appended drawings.
Brief Description of the Drawings
FIG. 1 is a representational view of a document page;
FIG. 2 is a representational view of a hierarchical tree structure resulting from a block selection operation on the document page illustrated in FIG. 1;
FIG. 3 is representational view of the document page which is created based on the hierarchical tree structure shown in FIG. 2;
FIG. 4 is a perspective view showing the outward appearance of an apparatus according to the present invention;
FIG. 5 is a block diagram of the FIG. 1 apparatus;
FIG. 6, comprising FIGS. 6A and 6B, is an example of an API library which includes both API functions and API data files which are utilized to create a page analysis program according to the present invention; and
FIG. 7 is a flow diagram describing a method for creating an application program utilizing the API of the present invention.
Detailed Description of the Preferred Embodiment
FIG. 4 is a view showing the outward appearance of a representative embodiment of the invention. Shown in FIG. 4 is computing equipment 10, such as a Macintosh or an IBM PC or PC compatible computer having a windowing environment, such as Microsoft Windows.RTM.. Provided with computing equipment 10 is display screen 12, such as a color monitor, keyboard 13 for entering user commands, and pointing device 14, such as a mouse for pointing to and for manipulating objects displayed on screen 12.
Computing equipment 10 includes a mass storage device such as computer disk 11 for storing data files which include document image files, in either compressed or uncompressed format, and for storing application program files which can include at least one block selection program and an application programming interface library which includes both functions and data files. Also stored in disk 11 are various hierarchical tree structure data for each document page which has been subjected to a block selection program, and various claimant page analysis programs for variously manipulating block selection functionality and for accessing data within the hierarchical tree structure.
Document image data is input by scanner 16 which scans documents or other images and provides bitmap image of those documents of computing equipment 10. The document image data may also be input into computing equipment 10 from a variety of other sources such as network interface 24 or other external devices via facsimile/modem interface 26. Printer 18 is provided for outputting process document images.
It should be understood that, although a programmable general purpose computer arrangement is shown in FIG. 4, a dedicated or stand alone computer or other type of data processing equipment can be used to practice the present invention.
FIG. 5 is a detailed block diagram showing the internal construction of computing equipment 10. As shown in FIG. 5, computing equipment 10 includes a central processing unit (CPU) 20 interfaced with computer bus 21. Also interfaced with computer bus 21 is scanner interface 22, printer for interface 23, network interface 24, fax/modem interface 26, display interface 27, keyboard interface 28, mouse interface 29, main memory 30, and disk 11.
Main memory 30 interfaces with computer bus 21 so as to provide random access memory storage for use by CPU 20 while executing stored program instructions such as the page analyzer API or various process steps of the block selection program. More specifically, CPU 20 loads those programs from disk 11 into main memory 30 and executes those stored programs out of main memory 30.
In accordance with a user's instructions, stored application programs provide for image processing and manipulating of data. For example, a desktop processing program, such as Wordperfect.RTM. for Windows, may be activated by an operator to create, manipulate, and view documents before and after block selection has been applied. Likewise, a page analysis program may be executed to run a block selection operation to analyze various types of image data in an input document page and to display the results of the page analysis to an operator via a windowing environment.
FIGS. 6A and 6B illustrate an API function library and an API data file, respectively, which are used to create a page analysis program according to the present invention.
To create a program for page analysis, the contents of the API library are retrieved from disk 11. In order to get the desired resulting program, the page analysis program is created by combining both functions and data files which are in the API library. For example, as shown in FIG. 6A, API function library 601 contains functions to manipulate processes of a block selection program and contains functions to access the data structures resulting therefrom. The API functions listed in API function library 601 are merely examples of some of the page analysis functions which can be utilized with the present invention. In this regard, other functions which could be used with the present invention are set forth in the appendix attached hereto.
As shown in FIG. 6A, API function library 601 contains page analysis functions 602, node relationship functions 609, text block functions 614, table block functions 618, line functions 623 and picture functions 625. A page analysis program could use some or all of these functions in order to perform a desired task. Each of these functions will be discussed in greater detail below.
In the present invention, page analysis functions 602 are utilized to manipulate various block selection process steps. Generally, prior to running page analysis, the page analysis program executes InitializePageAnalysis function 603 which initializes internal variables and allocates internal memory necessary for the block selection operations called by AnalyzePage function 604 to operate. Once the memory has been allocated and variables initialized, the page analysis program executes AnalyzePage function 604 which executes a page analysis operation on a stored input page of a document. The results of AnalyzePage function 604 are stored in memory as a hierarchical tree structure. At the completion of the AnalyzePage function 604, the hierarchical tree structure is completed and a root block ID is returned. If additional pages are to be analyzed, FreePageData is executed to clear memory space which was utilized in processing the previous page.
After page analysis is completed, the program executes EndPageAnalysis function 605 to free memory allocated by the InitializePageAnalysis function 603 and to exit the page analysis application.
Page analysis function 602 also includes BlockType function 606 which, when executed, returns an attribute of a specified block, such as picture, text, and non-text. PageSkew function 607, when executed, returns a skew angle of an entire page in tenths of a degree, wherein a 0 value indicates no page skew while, for example, a return value of 11 and 22 will represent a skew degree of 1.1 and 2.2, respectively.
Once the page has been analyzed and the hierarchical tree structure has been created, utilizing various functions in the API library, the user can interrogate the hierarchical tree structure in order to determine node relationships between nodes in the tree. That is, by utilizing the NodeRelationship functions 609, the user can interrogate a node in the tree to determine its relationship to another node in the tree. Thus, for example, GetAllChildren function 610 will return block IDs of all nodes of a specified type of image data having a specified parent or root node. Utilizing GetNumberOfChildren function 612, the program will return the number of children nodes of a specified node. Similarly, utilizing the GetParent function 613, the program will return a block ID of a parent node of a specified child node.
Other functions in API function library 601 permit the user to interrogate each of the nodes more closely. For example, utilizing various features of text block functions 614, a user can obtain a first text-line, a last text-line, or the next text-line of a text block. Utilizing GetFirstTextLine function 614, the program will return a pointer which is a block ID of a first text-line in a specified text node. Likewise, utilizing the GetLastTextLine function 615, and the GetNextTextLine function 617, the program will return a pointer which is a block ID or the last text-line and the next text-line, respectively, in a text block. If the user wishes to interrogate table nodes in the hierarchical tree, the user can examine data in each table node by utilizing table block function 618 which consists of GetFirstTableCell 619, GetLastTableCell 620, GetNextTableCell 621, and GetNumberOfCellsInTable function 622.
Other API functions such as line function 623 and picture function 625 permit the user to examine information in line nodes and picture nodes, respectively, in the hierarchical tree structure.
FIG. 6B illustrates an API data file which comprises parameters and definitions. The parameters and definitions contained in API data file 701 are utilized by associated API functions, defined in API function library 601. The API functions utilize their associated data files to access data in the hierarchical tree and to translate the accessed data into a user-friendly format. The parameters and definitions listed in the API data file 701 are merely examples of some of the definitions and parameters which can be utilized with the present invention. In this regard, other definitions and parameters which could be used with the present invention are set forth in the appendix attached hereto.
In the present invention, API data file 701 includes error codes 703 which are used by InitializePageAnalysis function 603, and AnalyzePage function 605 to return an error code in the case the page analysis cannot be performed due to the occurrence of a predefined event. For example, an error code may be output in the case the amount of memory required to perform a block selection operation is not sufficient or the skew of a page is too great to perform an accurate page analysis operation.
Block relationship-type definitions 704 are used by node relationship function 609 to return information regarding a relationship between specified nodes, such as parent, child, and no relation, etc. Line-type definitions 705 are accessed when returning information in response to line functions 623; and picture-type definitions 706 are accessed when returning information in response to picture functions 625.
Utilizing the API data file and API functions stored in API libraries 601 and 701, the user can create a page analysis program by combining at least one of the plural page analysis functions and at least one data file to create a desired API. For example, the user may select AnalyzePage function 603 in order to execute the block selection program stored on disk 11. Depending on the user's goal, the user may also select at least one of the plurality of relationship functions or one of the plurality of specific node feature functions, such as text block function, table block function, line block function, or picture block functions. Each of the selected functions will be coupled with its corresponding data file from API data file 701.
Upon combining desired functions and their associated data files, a page analysis program is formed. The newly created program can then be used to perform its predefined task. For example, the user's goal may be to perform optical character recognition (OCR) operation on all text blocks existing in the document page. Accordingly, the program could consist of AnalyzePage function 603 and GetAllChildrenBlock function 604 in order to return all children blocks of the root node which have a text attribute. By utilizing at least these two functions, the user can get all children blocks of the root node which are text and then submit the results of that program to an OCR processing program.
FIG. 7 is a detailed flow diagram of a method by which a page analysis program can be created utilizing the API library of the present invention.
In step S701, the API library, such as the one illustrated in FIGS. 6A and 6B, is downloaded from disk 11 into a random access memory (not shown). The API library loaded in random access memory is displayed to the user in step S702. Depending on the user's goal, one or more functions and corresponding data structures are selected and combined from the API library to obtain a program which produces a desired result.
Thus, in step S704, the user selects one or more functions and corresponding data files from the API library. In step S705, the functions and data files are compiled to form a desired application program.
In step S707, the newly created program is stored on disk 11 and can be executed from disk 11 to perform its desired function either directly by the user or by another program. ##SPC1##
Claims
What is claimed is:
1. An application programming interface for accessing page analysis functionality of a block selection program and for accessing data structures resulting therefrom, comprising: plural page analysis functionality which manipulate block selection functions, including functions to analyze a document page in order to identify text regions, non-text regions and bitmap image regions in the document page, to create a hierarchical tree structure corresponding to the analyzed document page, and to return a pointer to a root node of the analyzed page; and plural hierarchical tree structure accessing functions for traversing the hierarchical tree structure, and for locating desired nodes in the hierarchical tree structure.
2. An application programming interface according to claim 1, wherein the plural hierarchical tree structure accessing functions further include a function to return at least one pointer to at least one desired node.
3. An application programming interface according to claim 1, further comprising plural node interrogating functions for interrogating the at least one pointer to obtain attribute information of the at least one desired node corresponding thereto.
4. An application programming interface according to claim 3, wherein the plural node interrogating functions comprise text block interrogating functions, said text block interrogating functions including process steps to return a pointer to a first text line in a desired text block, to return a pointer to a last text line of the desired text block, return a pointer to a next text line in the desired text block, to return a number of text lines in the desired text block, to return a pointer to a previous text line in the desired text block, to return boundary information of a text line in the desired text block, and to extract a two-dimensional array of the bitmap image of the text line in the desired text block.
5. The application programming interface according to claim 3, wherein the plural node interrogating functions comprise picture block interrogating functions, said picture block interrogating functions including process steps to return information regarding picture type of desired picture block.
6. An application programming interface according to claim 3, further comprising a data file, said data file including definitions and parameters which are utilized by the plural page analysis functionality to manipulate block selection functions, by the plural hierarchical tree structure accessing functions to traverse the hierarchical tree, and to locate desired nodes, and by the plural node interrogating functions for interrogating the desired nodes.
7. An application programming interface according to claim 3, wherein the plural node interrogating functions comprise line block interrogating functions, said line block interrogating functions including process steps to return information regarding line type and other attributes of a desired line block.
8. An application programming interface according to claim 1, wherein the plural page analysis functionality comprises an initialization function to initialize the block selection program by setting internal variables and by allocating memory for processing and storing results thereof, an execution function for executing the block selection program on an input document image, and a termination function to free allocated memory set by the initialization function and to terminate the execution function.
9. An application programming interface according to claim 3, wherein the plural hierarchical tree structure accessing functions comprise a node relationship function which includes process steps for interrogating a relationship among nodes and for indicating how deeply nested a node is within the tree, a text node function which includes process steps for interrogating at least one node representing a text block, a table node function for interrogating at least one node representing a table block, a picture function for returning a type of picture represented by a picture node, and a line function for returning a type of line and other attributes represented by a line node.
10. An application programming interface according to claim 9, wherein the table node function includes process steps to return a pointer to a first table cell block of a specified table block, to return a table cell of a specified cell position in a table block, to return a pointer to a last table cell block of a specified table block, to return a next table cell block of a specified current table cell block, to return a number of table cell blocks of a specified block type of a table block, and to return a pointer to a previous table cell block of a specified current table cell block.
11. An application programming interface according to any of claim 2-4, wherein the returned pointer is a block identification.
12. A method of utilizing a page analysis API to create a page analysis program for accessing page analysis functionality of a block selection program and for accessing data structure resulting therefrom, the method comprising the steps of: storing, in memory, plural page analysis functionality which manipulate block selection functions, including functions to analyze a document page in order to identify text regions, non-text regions and bitmap image regions in the document page, to create a hierarchical tree structure corresponding to the analyzed document page, and to return a pointer to a root node of the hierarchical tree structure which represents the analyzed page, and also storing in memory plural hierarchical tree structure accessing functions for traversing the tree and for locating desired nodes in the tree; selecting, from memory, at least one of the plural page analysis functionality and at least one of the plural hierarchical tree structure accessing functions; and combining the at least one of the plural page analysis functionality selected and the at least one of the plural hierarchical tree structure accessing functions selected so as to create the page analysis program, whereby the created page analysis program is capable of accessing page analysis functionality of a block selection program and is capable of accessing a hierarchical tree structure resulting therefrom.
13. A method of utilizing a page analysis API to create a page analysis program according to claim 12, wherein the created page analysis program is also capable of returning at least one pointer to at least one desired node.
14. A method of utilizing a page analysis API to create a page analysis program according to claim 12, further comprising the steps of storing, in memory, plural data files which include data files which are associated with the plural page analysis functionality, and the plural hierarchical tree accessing functions, and combining at least one of the plural data files associated with the at least one selected plural page analysis functionality and combining at least one of the plural data files associated with the at least one selected plural hierarchical tree structure accessing function.
15. An application programming interface for accessing page analysis functionality of a block selection program and for accessing data structures resulting therefrom, comprising: plural page analysis functionality which manipulate block selection functions including functions to analyze a document page, to create a hierarchical tree structure corresponding to the analyzed document page, and to return a pointer to a root node of the analyzed page; plural hierarchical tree structure accessing functions for traversing the hierarchical tree structure, and for locating desired nodes in the hierarchical tree structure; and plural node interrogating functions for interrogating the at least one pointer to obtain attribute information of the at least one desired node corresponding thereto, wherein the plural node interrogating functions comprise text block interrogating functions, said text block interrogating functions including process steps to return a pointer to a first text line in a desired text block, to return a pointer to a last text line of the desired text block, return a pointer to a next text line in the desired text block, to return a number of text lines in the desired text block, to return a pointer to a previous text line in the desired text block, to return boundary information of a text line in the desired text block, and to extract a two-dimensional array of the bitmap image of the text line in the desired text block.
16. An application programming interface for accessing page analysis functionality of a block selection program and for accessing data structures resulting therefrom, comprising: plural page analysis functionality which manipulate block selection functions including functions to analyze a document page, to create a hierarchical tree structure corresponding to the analyzed document page, and to return a pointer to a root node of the analyzed page; plural hierarchical tree structure accessing functions for traversing the hierarchical tree structure, and for locating desired nodes in the hierarchical tree structure; and plural node interrogating functions for interrogating the at least one pointer to obtain attribute information of the at least one desired node corresponding thereto, wherein the plural node interrogating functions comprise picture block interrogating functions, said picture block interrogating functions including process steps to return information regarding picture type of desired picture block.
17. An application programming interface for accessing page analysis functionality of a block selection program and for accessing data structures resulting therefrom, comprising: plural page analysis functionality which manipulate block selection functions including functions to analyze a document page, to create a hierarchical tree structure corresponding to the analyzed document page, and to return a pointer to a root node of the analyzed page; plural hierarchical tree structure accessing functions for traversing the hierarchical tree structure, and for locating desired nodes in the hierarchical tree structure; and plural node interrogating functions for interrogating the at least one pointer to obtain attribute information of the at least one desired node corresponding thereto, wherein the plural hierarchical tree structure accessing functions comprise a node relationship function which includes process steps for interrogating a relationship among nodes and for indicating how deeply nested a node is within the tree, a text node function which includes process steps for interrogating at least one node representing a text block, a table node function for interrogating at least one node representing a table block, a picture function for returning a type of picture represented by a picture node, and a line function for returning a type of line and other attributes represented by a line node.
Patent Citations (4)
Non-Patent Literature (3)
- Fein, et al., "Model-based control strategy for document image analysis", Proceedings--SPIE-The International Society for Optical Engineering, vol. 1661, Feb. 10-12, 1992, pp. 247-256.
- Yamada, et al., "Mixed Mode Document Processing System", IEE/IEICE Global Telecommunications Conference 1987, Nov. 15-18, 1987, pp. 30.2.1-30.2.5.
- Tsujimoto, et al., "Understanding Multi-articled Documents", 10th International Conference on Pattern Recognition, Jun. 16-21, 1990, pp. 551-556.