JPH0498582A - Image generator - Google Patents

Abstract

PURPOSE:To obtain a device easy to use by recognizing a content by reading a document original in which a character is written on a prescribed form by handwriting, and automatically making a recognized content into a graph. CONSTITUTION:The document original for input is generated by writing a numeric character on a blank part blocked with a frame line by handwriting. A scanner 102 reads the document original placed on an original platen 101. Thence, a frame is recognized at a frame recognition part 106, and the character is segmented frame by frame. The frame line is separated from the character by segmented information. A character recognition part 108 performs the matching of a handwritten character with a dictionary, and specifies the character as the numeric character. When no specification is performed, recognition is performed by performing the structure analysis of the contour of the numeric character. Such processing is applied to all characters. A recognition result is stored in memory 109. Furthermore, the graph is generated at an image generating part consisting of a numeric value calculation part 111 and a image generation processing part 112, and it is stored in memory 113 as image data.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to an image generation device such as a copying machine, a facsimile machine, etc., which is equipped with a character recognition function and a drawing function.

[Conventional technology]

Conventionally, devices that optically read a document, perform various processing on the read image, and then generate a new image include, for example, ■Japanese Patent Application Laid-Open No. 58-85670, ■Japanese Patent Application Laid-Open No. 54-1187720. There are some disclosures in each publication.

Below, they will be explained in order.

■The image generating device disclosed in Japanese Patent Application Laid-Open No. 58-85670 is related to a device that prints out an image combined on gift paper, and the image is written in a predetermined format with relevant parts marked. Reads the application form, determines the type of gift desired based on the mark, reads the corresponding term pattern from the memory, and matches the term pattern and the sender's name pattern written on the application form to the positions to be entered. The pattern is then attached and combined, and the combined pattern is printed on gift paper.

■The facsimile disclosed in Japanese Patent Application Laid-open No. 54-118720 has a configuration convenient for transmitting and outputting graphs, and the sending side scans the document two-dimensionally and optically reads the data. A graph can be generated on the receiving side simply by encoding, extracting the codes and quantities for the graph items, and transmitting them.

[Problem to be solved by the invention]

However, the techniques disclosed in the above publications have the following problems.

■The image generation device disclosed in Japanese Patent Application Laid-open No. 58-85670 allows easy selection of gift patterns, and
Although it can be easily combined with the sender's name, since it is determined by reading the mark, there is a natural limit to the number of patterns that can be stored, and the range of use is extremely restricted. Also,
Since the sender's name is also simply a copy, it is impossible to perform any secondary value-added processing using recognition means.

■In the facsimile disclosed in Japanese Patent Application Laid-Open No. 54-118720, the transmission capacity is reduced through encoding, making it possible to use the communication channel effectively, and also shortening the transmission time, but basically the graph Since the format is faithfully transmitted, it is not possible to recognize the contents of the graph and convert it to another format.

The present invention has been made in view of the above, and it recognizes the contents of a document read optically, graphs the contents in various formats to increase added value, and automatically generates data based on the recognized contents. The aim is to make graphs easy for anyone to use by creating graphs.

Another object of the present invention is to enable a scribe to write characters on a document in accordance with his or her daily habits without being conscious of whether they are right-aligned or left-aligned, thereby improving the user's ability to use the device.

[Means to solve the problem]

In order to achieve the above object, the present invention includes a reading means for reading a form document in which characters are handwritten in a predetermined format, a storage means for storing the content read by the reading means as image data, and a storage means for storing the content read by the reading means as image data. frame line recognition means for recognizing a frame line in image data; separation means for separating the frame line and character portion recognized by the frame line recognition means; and recognition of the character portion separated from the frame line by the separation means. a display means for displaying the recognition result by the recognition means, a storage means for storing the recognition result by the recognition means, an instruction means for instructing a storage location of the recognition result by the recognition means, and the recognition means a correction means for correcting the recognition content as necessary; a determination means for determining whether the characters written on the document are right-justified or left-justified; a calculation means for performing numerical calculations according to the recognition content; a drawing means for drawing the recognition content as a graph based on the calculation result by the calculation means; and printing the composite of the graph drawn by the drawing means and the recognition content. an image output means for specifying a type of graph to be drawn by the drawing means; and an interrupt means for causing other image generation processing to be executed while the recognition means and the drawing means are performing recognition and drawing processing. The present invention provides an image generation device equipped with the following.

Further, it is preferable that the correction means has an input means for inputting a correction point and a correction value from a display screen based on the recognition content displayed on the display means.

Furthermore, it is preferable that the determining means determine whether the characters written on the form manuscript are right-aligned or left-aligned by comparing the total sum of blanks at the right end and left end of each number column entry field.

[For production]

According to the image forming apparatus of the present invention, a frame line is recognized from image data obtained by reading a document document in which numbers have been written by hand, the frame line and a character part are separated, and the separated character part is recognized. The recognition result is displayed on the display means, and the recognition content is corrected as necessary. In addition, it determines whether the characters written on the document are right-aligned or left-aligned, performs numerical calculations according to the recognition content based on the determination results, and draws the recognition content as a graph based on the calculation results. , the drawn graph and the recognized content are combined and printed.

Furthermore, based on the recognized content displayed on the display means, the recognized content is corrected by inputting the correction location and correction value from the display screen.

〔Example〕

Hereinafter, one embodiment of the present invention will be described in detail using the drawings.

FIG. 1 is a block diagram showing the configuration of an image generation device according to an embodiment of the present invention.

The image generation device in this embodiment includes a form document 2 for input.
01 (see FIG. 2), a scanner 102 that optically reads the form document 201 placed on the document table 101, and converts the electric signal from the scanner 102 into a digital signal. An A/D conversion unit 103, a memory 104 that stores the digital signal converted by the A/D conversion unit 103 as image data, and information regarding the frame configuration is recognized from the image data stored in the memory 104. A frame line recognition unit 106 , a character cutting unit 107 that cuts out the text written in the frame one frame at a time based on the information regarding the frame recognized by the frame line recognition unit 106 , and a character cutting unit 107 that cuts out the text written in the frame one frame at a time. An OCR unit 105 consisting of a character recognition unit 108 that recognizes issued characters, and a 3ocR unit 1
A memory 109 for storing recognition results by 054m, and an OC
A drawing section 110 includes a numerical calculation section 111 and a drawing processing section 112 that create a graph based on the numbers identified in the R section 105, and a memory 113 that stores the graph created by the drawing section 110 as image data. ,
A switch 118 that switches the flow of data between a printing unit 115 that executes image printing processing, a paper ejection unit 116 that outputs paper printed by the printing unit 115, and a switch 118 that switches the flow of data.
, a touch panel 2 operation section 123 that has a display section 120 that displays the read contents of the document document, and that performs drawing commands, mode settings, correction of recognition results, etc., and a buffer 131 that temporarily stores processing data. , a switch 132 that controls the flow of data and interrupts normal image generation processing while recognition/drawing processing is being performed, and memory 109
an address calculation unit 134 that calculates the address of the form document 201; and an in-memory search unit 135 that determines whether the numbers written in each item of the document document 201 are written right-aligned or left-aligned. It is composed of.

Further, FIG. 2 shows the structure of a form document 201 for input,
The document document 201 is composed of a frame line 202 as shown in the figure, and a blank 2 partitioned by the frame tili 202.
Numbers are written in 03 by hand.

FIG. 3 is a diagram showing the configuration of the display section 120 in the operation section 123, including a recognition content display section 301 that displays the recognition result of the document document 201 in the same format as the document, and the contents of the command input from the operation section 123. an instruction display section 302 that displays
It consists of a numeric keypad display section 303 that displays numbers to be input. To enter numbers, use the numeric keypad display section 30 above.
This is done by touching the number 3.

In FIG. 1, a document document 201 for input on which necessary information has been written by hand is placed on a document table 101. As shown in FIG.

The document document 201 is read by the scanner unit 102 and converted into an electrical signal, and then converted into a digital signal by the A/D conversion unit 103 and stored as image data.

The image signal stored in the memory 104 is recognized as a number in the 008 section 105, and a drawing section 110 creates a rough image based on the number identified in the 008 section 105.

The recognition processing in the 008 unit 105 will be described below with reference to the flowchart in FIG.

First, the frame line recognition unit 106 extracts line segment elements of the frame 9202 from the image data stored in the memory 104, and if the interval between the elements is less than a certain dot, the elements are connected to form a line segment. If the distance between the extracted line segment and the adjacent line segment is less than a certain darkness value, the line segments are connected to form a frame. If the size of the constructed frame is larger than a certain darkness value, it is recognized as a frame, and if it is smaller, it is not recognized as a frame (step 5401
).

Next, information regarding the structure of the recognized frame, such as the number and position of the frame, is identified and extracted. Information regarding each extracted frame line 202 is stored in memory 135. At this time, first, the first address of the storage destination memory 135 is designated, and information about the frame line 202 on the memory 104 is sequentially written from that address (step 5402).

5A and 5B show the structure of information regarding frames stored in the memory 135.

As shown in the figure, the number of the recognized frame, the position in the X direction (horizontal direction), and the position in the Y direction (vertical direction) of each frame are stored.

To specify the position of the frame, select the starting point of the frame in the X and Y directions (xs, ys
) and the end point (xe, ye) in the x and y directions.

Based on the information regarding these frames, the character cutting unit 107 cuts out the numerical part written in each frame one frame at a time from the image data stored in the memory 104 (step 540
3). Then, projections are taken in the vertical and horizontal directions from the information for one cut out frame, and the frame lines and numbers are separated from the projection state (step 5404).

These numeric images cut out and separated one frame at a time are matched with a dictionary in the character recognition unit 108 (step 5405), and it is determined whether recognition is specified as a numeral (step 5406). In step 5406, if recognition is not specified by matching with the dictionary, the structure of the outline of the number is analyzed and recognized (step 5407). The character recognition process described above is repeated until all characters are completed (step 540B). At this time, if the number written in the frame is deemed unrecognizable because its shape is inappropriate or obscured by noise, etc., a reject symbol indicating that it is unrecognizable will be added. If no characters are written within the cut-out frame, a blank symbol indicating that the frame is blank is used as a recognition result instead of a number.

The recognition results in the character recognition unit 108 are stored in the memory 109. The starting address of the storage destination memory 109 is specified, and the recognition results are sequentially written from that address.

The contents stored in the memory 109 are displayed on the display section 120 in the operation section 123 through the data bus 140 in the same format as the form. The address calculation unit 134 determines the storage destination address of the memory 109 and the position of the frame on the form corresponding to the address, and stores the memory 109 at the determined position.
Display the contents of the corresponding address above. The displayed recognition result is compared with the document document 201, the erroneously recognized portions and rejected portions are touched with a fingertip, and the correct values are re-inputted by touching the numbers displayed on the numeric keypad display section 303.

In FIG. 6, memory 1 is displayed in form on the display unit 120.
The contents of 09 are shown below.

Here, an example is shown in which the number 7 in location 501 is incorrectly recognized as 9, and the number 2 in location 502 is rejected due to unrecognizability.

FIG. 7 shows a flowchart of the process of correcting the recognition results stored in the memory 109.

By pointing to the correction input command on the display unit 120,
The address counter sets the starting address of the memory 109 in which the recognition result is stored (step).

5701).

The user specifies the location to be corrected by touching the display unit 120, and inputs the coordinates of the changed location (step 5702).

Then, the address calculation unit 134 calculates which frame is specified from the row and column of the specified frame (step 5703), and moves the address counter by that number from the first address (step 5704).

FIG. 8 shows the relationship between the start address in the memory 109, the relative value of coordinates, and the address counter.

A correction value is input from the numeric keypad display section 303 of the display section 120 (step 5705), and the content of the address specified by the address counter is rewritten with the input value (step 5706). The corrected input value is transferred to the memory 109 through the lotus 140. If it is determined that the correction has not been completed and the correction is to be repeated, the above flow is restarted from inputting the coordinates of another changed location (step 5702).
step 5707).

Next, the graph drawing process will be explained with reference to the flowchart of FIG. 9 showing the relationship with the recognition result correction process.

First, it is determined whether or not there is a reject in the recognition result (step 5901). If there is a reject, the display unit 120
The recognition result is displayed on the screen (step 5902), the correction location is specified and the correction value is input (step 5903), and the contents of the memory 109 are corrected (step 5904). Next, it is determined whether the correction has been completed (step 5910).
, if it is determined that the process has ended, the flow proceeds to step 590.
If it is determined that the correction has not been completed, the flow repeats the operations from step 5903 until it is determined that the correction has been completed in step 5910.

On the other hand, if there is no reject in step 5901, it is determined whether the drawing mode is set (step 590
5). Here, if there is an erroneous recognition in the recognition result, the drawing mode is not selected, and if it is determined that there is no drawing mode as a result, steps 5902 to 59
04 correction processing is executed.

When the drawing mode is selected, it is determined in step 5905 that the drawing mode is selected, and the following drawing processing is executed.

A graph drawing command based on the result of the recognition of the numbers is transmitted from the operation unit 123 via the control bus 141. This instruction causes the memory 10 to
The contents of 9 are transferred to the drawing section 110.

The drawing unit 110 determines the maximum value, minimum value, and interval of the scale necessary to create a graph based on the results of recognition items related to drawing stored in the memory 109, or the percentage of each item in the entire graph and drawing. The numerical calculation unit 111 calculates the target area and the like.

The drawing processing unit 112 creates a graph using different drawing patterns based on the calculation results of the numerical calculation unit 111 so that the area corresponding to each item can be distinguished from other item areas (step 5906).

The graph created by the drawing processing unit 112 is stored in the memory 113 as image data.

The image data stored in the creation memory 113 by the OCR unit 105 and the drawing unit 110 is transferred to the bus 146 and then printed through the print unit 115 (step 5907), resulting in a form with handwritten numbers input from the manuscript table 101. The original 201 is delivered as a graph to the paper output unit 116.
(Step 5909). At this time, along with the graph, the contents used for drawing, which are stored in the memory 109, are also written to the memory 113 via the bus 143 and printed (step 5908). Print section 115
FIG. 10 shows an example in which the graph and the contents of the memory 109 are printed out.

The corrected contents of the memory 109 are retained and the control lot 1
The graph can be drawn as many times as desired depending on the type of graph to be created, which is input through the operation unit 123 through the screen 42.

Further, initialization of the memory 109 is controlled by a control lot 140. The address calculation unit 134 sets the address counter to the first address of the recognition result stored in the memory 109, and sequentially clears the contents indicated by the address.

The recognition and drawing functions of the OCR section 105 and drawing section 110 are as follows:
The system is separate from the normal digital image generation function, and while the recognition/drawing process is being performed, the normal digital image generation process can also be executed in parallel via the data bus 144 by interrupt. Control lotus 147 is controlled by command input from operation unit 123.
The data flow is controlled by switching the switch 132 through the switch 132. Furthermore, the data bus 144 is
and Hass 145 to avoid data conflicts between the two systems.

Figure 11 shows two cases of normal image generation and recognition/drawing processing.
Show the path of the lineage.

FIG. 12 shows a flowchart regarding interruption of normal image generation processing during recognition/drawing processing.

When the recognition/drawing mode is set from the operation unit 123 through the lotus 147, the switch 132 closes the normal path of the data bus 1440 and opens the circuit to the OCR unit 105.
(Step 51201). This makes the input data OC
The image flows to the R section 105 and the drawing section 110, where the above-mentioned recognition and drawing processing is performed (step 5I202).

Next, it is determined whether or not there is an interruption in the normal function (step 51203). In this state, when a command to perform normal image generation processing is input from the operation unit 123, the lotus 147
The switch 132 opens the circuit to the data bus 144 and cuts off the circuit to the OCR section 105 (step 51).
204). When data is input in this state (step 51205), the input data flows to normal image generation processing via the data north 144 (step 51206).
, normal image generation processing is executed by interruption (step 51207).

In the interrupt state, the data of data north 144 flows to lotus 146 until the data is transmitted to lotus 145. Next, it is determined whether the data has been transmitted to the lotus 145 (step 51208), and in this state, the lotus 1
45 (If it is determined that data has not been transmitted to the lotus 145, step 5
1205 is repeated until it is determined that it has been transmitted in step 81208), and the data bus 144
The data on the bus 145 is varnished to the buffer 131 (step 51209), and the data on the bus 145 is transmitted to the bus 146 (step 51210).

When all data from bus 145 is transferred to bus 146 (step 51211), buffer 1 is transferred to bus 146.
The data in 31 is transmitted (step 51212).

Next, FIG. 3A shows a flowchart for detecting blanks present at both ends of the item field for numerical entry and determining whether it is right-aligned or left-aligned.

The counting counters LC and RC are cleared to zero (step 51301). After that, enter the first address of the memory 109 in the address counter (step 51302),
Blanks at the beginning and end of the fixed-length character string range are searched for (step 51303).

After searching for a blank, it is determined whether there is a blank on the left end or right end (Step 51304)
, if it is determined that there is a blank on the right edge, left justified L
Increment C (step 51305). On the other hand, if it is determined that there is a blank on the left edge, right-justify R
Increment C (step 51306).

After that, step 51307 to access the next string.
), it is determined whether all characters have been completed (step 31
308). If it is determined that all the characters have not been completed, the process returns to step 51303 and the following process is repeated until it is determined that all the characters have been completed in step 51308. On the other hand, if it is determined that all the characters have been filled, the majority vote of LC and RC determines whether the characters are right-aligned or left-aligned on the document document 201 (step 51309).
.

FIG. 13B is a flowchart showing the operation of determining right alignment or left alignment of characters on the document document 102 from the sum of RC and LC for all characters (step 51310), and determining whether RC≧LC. (Step 51311).

Here, if it is determined that RC≧LC is not satisfied, it is determined that left alignment is performed, and if it is determined that RC≧LC, it is determined that right alignment is performed.

FIG. 14 shows examples of number sequences written in right-justified format, number sequences written in left-justified format, etc.

Here, when all characters are blank (FIG. 14(a)), the counters that increase are RC and LC, and it is determined that the characters are right-justified. Further, in the case of FIG. 14(b), the increasing counter is determined to be right-aligned by RC.

In the case of FIG. 14(C), the increasing counter is determined to be left-justified by LC.

Furthermore, in the case of FIG. 14(d) in which all the characters are filled, there is no incrementing counter, and it is determined that the characters are right-justified.

In the case of the character shown in FIG. 14(e) whose purpose is unknown, the incrementing counter is considered to be right justified in RC and LC, and the blank at the right end is replaced with 0.

〔Effect of the invention〕

As described above, according to the image generation device according to the present invention, handwritten numerical data can be output as a graph that can be intuitively understood, and the type of graph can also be selected according to the data. It can be expressed in an appropriate form depending on the content and nature of the data, and since numerical data is also displayed together with the graph, objective data can be shown along with intuitive images.

Since graph output is achieved by simply handwriting numerical data, anyone can easily graph the necessary data.

In addition, since numerical data recognition and graph drawing can be performed in parallel with the normal image generation function, normal image generation processing can be performed while recognition and drawing processing is being performed, improving office efficiency. .

According to the invention of claim 2, the correction means has means for inputting the correction part and correction value from the display screen based on the recognition content displayed on the display means, so that comparison with the original can be easily performed. , errors in recognized content can be quickly corrected with simple operations.

According to the third aspect of the present invention, the scribe can write numbers in accordance with his daily habits without being conscious of right-justification or left-justification when writing numbers on a document, thereby improving the usability of the device.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an image generation device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram showing the configuration of a document for long use, and FIG. 3 is a configuration of the display section in the operation section. 4 is a flowchart showing the recognition process in the OCR section, FIGS. 5A and B are explanatory diagrams showing the structure of information related to frames stored in memory, and FIG. 6 is a flowchart showing the recognition process in the OCR unit. An explanatory diagram showing the contents of the displayed memory, Fig. 7 is a flowchart showing the process of correcting the recognition result stored in the memory, and Fig. 8 shows the relationship between the start address and relative values of coordinates and address counters in the memory. 9 is a flowchart showing the relationship between the graph drawing process and the recognition result correction process. FIG. 10 is an explanatory diagram showing an example of printing out the graph and the contents of the memory by the print unit. Fig. 12 is an explanatory diagram showing two routes for normal image generation processing and recognition/drawing processing; Figure 13A is a flowchart showing the operation of detecting blanks present at both ends of the item field for numerical entry and determining whether it is right-aligned or left-aligned.Figure 13B is the RC for all characters. L
Flowchart showing the operation of determining right alignment or left alignment of characters on a document document from the sum of C, FIG. 14(a)
-(e) are explanatory diagrams showing examples of a number sequence written in right-justified form, a number sequence written in left-justified form, etc., and showing incrementing counters and determination results for each. Explanation of symbols 104゜112・Scanner 109 113--Memory OCR section 106--Frame line recognition section Character cutting section 108-Character recognition section Drawing section 111-Numerical calculation section Drawing processing section 115-Print section 132--Switch display Section 123 Buffer 134 Memory search section Form manuscript operation section Address calculation section

Claims (3)

[Claims] (1) A reading means for reading a document document with handwritten characters in a predetermined format, a storage means for storing the content read by the reading means as image data, and recognizing a frame line in the image data. a frame line recognition means; a separation means for separating the frame line and the character part recognized by the frame line recognition means; a recognition means for recognizing the character part separated from the frame line by the separation means; display means for displaying recognition results; storage means for storing recognition results by the recognition means; instruction means for instructing where to store the recognition results by the recognition means; correcting content recognized by the recognition means as necessary. a determining means for determining whether the characters written on the document document are right-justified or left-justified; and a numerical calculation according to the recognition content by the recognition means based on the result by the determining means. a calculation means; a drawing means for drawing the recognized content as a graph based on the calculation result by the calculation means; an image output means for combining and printing the graph drawn by the drawing means and the recognized content; and the drawing means. An image generation method comprising: a specifying means for specifying the type of graph desired to be drawn; and an interrupt means for causing other image generation processing to be executed while the recognition means and the drawing means are performing the recognition and drawing processing. Device. (2) The image generation device according to claim 1, wherein the correction means has an input means for inputting a correction part and a correction value from a display screen based on the recognition content displayed on the display means. . (3) In claim 1, the determining means determines whether the characters written on the form manuscript are right-aligned or left-aligned by comparing the sum of blanks at the right end and the left end of each number column entry field. An image generation device characterized by: