JPH04117058A - Picture processor - Google Patents

Abstract

PURPOSE:To describe information whose recognition resolution is different on same input paper by discriminating a describing area and an information kind of specific picture information, detecting an area of a recognition object based on an information kind mark so as to match the resolution detected by a facsimile interface section with the resolution required for a recognition section. CONSTITUTION:The picture processor is provided with an information area detection section 103 detecting the area of recognition object based on an information kind mark and a resolution conversion section 104 to match the resolution detected by a facsimile interface section 101 with the resolution required for a recognition section 103. When the information of plural kinds in the input paper received from a facsimile equipment is recognized, the resolution at the reception and the information kind mark are detected and resolution is converted to match the resolution of various recognition information with the resolution of recognition sections 105,106 based on them and each information is recognized, Thus, even when the recognition sections 105,106 do not have plural recognition parameters and recognition dictionaries, plural sets of information having different resolution for recognition processing is described on one page.

Description

[Detailed Description of the Invention] [Industrial Application Field 1] The present invention detects information type marks indicating various types of image information in input paper received from a facsimile, etc., detects the resolution at the time of reception, and detects the resolution of this information. The present invention relates to an image processing device that performs resolution conversion of image signals of various image information entry areas based on , and performs recognition processing of various information.

[Prior art] Conventionally, mark recognition in image signals received from facsimile machines, etc. has a recognition parameter (threshold value, etc.) for each resolution of the input device to correspond to the resolution at the time of input, because the recognition process is simple. It was recognized using the recognized recognition parameters.

[Problem to be solved by the invention] However, recognition dictionaries for character recognition are complex, and it is extremely difficult to create a recognition dictionary for each input device resolution.
The resolution of the input device is equivalent to the high quality mode of G3 facsimile. Therefore, if marks and characters are allowed to coexist on one page, mark recognition may have a recognition parameter of the resolution used for character recognition, and both may be recognized at the resolution of character recognition. However, when inputting graphic information at 400 ppi on a G4 facsimile, characters cannot be mixed on one page, so the input is divided into two pages for processing.

Furthermore, if a new resolution recognition process is added in the future, that process will be added to a separate page, and even if the amount of information can be contained in one page, multiple pages of input will be required, increasing communication time. There is a problem.

In order to eliminate these problems, this invention detects an information type mark from an image signal received from a facsimile, etc., determines the information type, and specifies the information type mark based on the resolution and information type at the time of reception. The resolution of each entry area is converted so that each recognition section can perform recognition processing.The purpose is that each recognition section can process multiple recognition parameters,
The purpose is to allow multiple pieces of information with different recognition processing resolutions to be written on one page, even without a recognition dictionary, and to facilitate the addition of new recognition processing.

[Means for Solving the Problems] The image processing device according to the present invention determines the entry area and information type of specific image information, and detects the area to be recognized based on the information type. The apparatus includes an area detection section and a resolution conversion section for matching the resolution detected by the facsimile interface section with the resolution required by the recognition section.

[Operation J] In this invention, when recognizing multiple types of information in an input form received from a facsimile, the resolution and information type mark at the time of reception are detected, and based on these, various types of recognition information are sent to each recognition unit. This allows each information to be recognized by performing resolution conversion to match the resolution, and there is no need for each recognition unit to have processing for multiple resolutions as in the past. There is no need to input information that cannot be recognized at different resolutions on a separate page, as is the case with 400ppi facsimile. Furthermore, even if a recognition device with a new resolution is added, there is no need to modify each conventional recognition unit by adding a resolution conversion function. Furthermore, when entering input paper via facsimile, there is no restriction on resolution except in special cases (information that cannot be recognized by registering graphic information or converting resolution).

[Embodiment] Before explaining the embodiment shown in FIG. 1, input using an input paper will be explained with reference to FIGS. 2 and 3.

FIG. 2 is a diagram showing an example of an input form used in this invention, where 1 is an input form, 2 is a reference position mark indicating the left edge of this input form 1, and 3 is the upper edge of this input form 1 or the entire information to be filled in. 4 is also an end mark that indicates the bottom edge of input paper 1 or the bottom edge of the entire entry information; 5 and 6 are mark line designation marks that indicate the position of the mark line in the sub-scanning direction; 7, 8 is a character line designation mark that indicates the position of the character line in the sub-scanning direction, 9 and 10 are graphic line designation marks that indicate the position of the graphic line in the sub-scanning direction, 11 is a mark entry field for entering a mark, and 12 is a character line designation mark. Text entry field to fill in,
Reference numeral 13 is a figure entry column in which an arbitrary figure is entered. In addition,
Reference position mark 2, start mark 3, end mark 4
, mark line designation marks 5, 6, character line designation marks 7.8
, graphic line designation marks 9 and 1o are called control marks, mark line designation marks 5 and 6, character line designation marks 7 and 8, and graphic line designation marks 9 and 10 are called information type designation marks, which are detected by the facsimile sensor. Write in the colors available. Also,
The mark entry field 119, the character entry field 12, and the figure entry field 13 are written in a dropout color that cannot be detected by a facsimile sensor.

The size of these control marks is predefined, and the position in the main scanning direction from the reference position mark 2 is also defined (in addition, mark entry field 11, character entry field 12
2. The size and arrangement of the figure entry field 13 are also determined.

However, the size of the graphic entry field 13 in the sub-scanning direction is arbitrary.

Figure 3 shows multiple mark entry fields 11 from various control marks;
2 is a diagram showing a method of detecting an information entry area in which a plurality of character entry fields 12 or a graphic entry field 13 are written, 20 is a mark entry field area in which a plurality of mark entry fields 11 are written, and 21 is a plurality of character entry fields 12. 22 is a graphic entry field area in which a graphic entry field 13 is located.

Next, a method for detecting an information entry area will be explained using FIG. First, when the input paper 1 is input, it is stored on the image signal memory in the form shown in FIG. Then, the image signal in the image signal memory is checked, and the image signal of the reference position mark 2 which has a specified width and length and is present at the left end of the input paper 1 is detected. When the image signal of the reference position mark 2 is detected, the image signal of the start mark 3 located in a specific range from the upper end of the image signal of the reference position mark 2 to the lower right is detected. When the image signal of the start mark 3 is detected, the information type designation mark is detected.

First, below the start mark 3 and at a specific distance to the right from the reference position mark 2, a mark line designation mark 51, a character line designation mark 7, a graphic row designation mark 9, and a specific width exist.
．． An image signal of any of the end marks 4 is detected. When the image signal of the mark row designation mark 5 is detected, the image signal of the mark row designation mark 6 at a specific position including the allowable inclination from the position of the reference position mark 2 indicating the upper end of the mark row designation mark 5 is detected. When the image signal of the character line designation mark 7 is detected, the image signal of the character line designation mark 8 at a specific position including the allowable inclination from the position of the reference position mark 2 indicating the upper end of the character line designation mark 7 is detected. When the image signal of the figure row designation mark 9 is detected, the image signal of the figure row designation mark 1o located at a specific position considering the allowable inclination from the position of the reference position mark 2 indicating the upper end of the figure row designation mark 9 is detected. do. When the marked line designation mark 5.6 is detected, the mark entry field area 2 surrounded by the marked line designation mark 6 starts from the right side of the marked row designation mark 5 and is a certain distance away from the reference position mark 2.゜, character line designation mark 7.
8 is detected, the character entry field area 21 surrounded by the character line designation mark 8 is moved from the right side of the character line designation mark 7 at a certain distance from the reference position mark 2 to the graphic row designation mark 9, If 10 is detected, the figure row designation mark 9
A figure entry field area 22 surrounded by figure line designation marks 10 from a place a certain length away from the reference position mark 2 on the right side of
Detect. Finally, the image signal of the end mark 4 is detected to complete the detection of the information entry area.

FIG. 4 is a diagram showing an example of a conversion ratio when performing resolution conversion.

Next, a method for determining the resolution conversion ratio will be explained using FIG. Mark recognition can process only the resolution for G3 facsimile standard mode input, character recognition can process only the G3 facsimile high quality mode input resolution, and the graphic entry field area 22 can only process the resolution for input in facsimile mode. An explanation will be given based on an example in which processing is performed.

At this time, mark entry field area 205 character entry field area 21
, when the input form 1 with the figure entry area 22 is input in the standard mode of 63 facsimile, the resolution conversion is no resolution conversion for the mark entry area 20, and sub-scanning for the character entry area 21. Only the direction is subjected to double resolution conversion, and the graphic entry field area 22 is not subjected to resolution conversion. When the input paper 1 is input in high quality mode of G3 facsimile or 200 ppi of G4 facsimile, the resolution conversion is as follows: For mark entry area 20, sub-scanning is converted to 1/2 resolution, and for character entry area 21, the resolution is converted to 1/2. There is no resolution conversion for the graphic entry field area 22. Input paper 1
4 When inputting at 400ppi facsimile, the resolution conversion is to change the main scanning to 1/1 for the mark entry area 20.
2x resolution and 1/4x resolution in the sub-scanning direction, 1/2x resolution conversion in both the main and sub-scanning directions for the text entry area 21, and no resolution conversion for the graphic entry area 22. .

As described above, the resolution conversion ratio is determined based on the input resolution of the facsimile and the type of information.

FIG. 1 is a block diagram showing one embodiment of the present invention,
An example will be shown in which input form 1 input from a facsimile is processed and sent to a computer. 100 is a facsimile machine for inputting the input paper 1; 101 is a facsimile interface unit that terminates the facsimile procedure and extracts an image signal from the facsimile signal; 102 is an image signal memory unit that stores image signals; An information area detection unit determines the entry area and information type of various information based on various control marks, and 104 is a resolution conversion unit that converts the resolution of the image signal in the specified area of the image signal memory at a specified ratio. , 105 is a mark recognition unit that recognizes the mark entry field 11 written in a specific area of the mark entry field area 20;
106 is a character recognition unit that recognizes the character entry field 12 written in a specific area of the character entry field area 21; 107 is a figure extraction unit for the image signal of the figure entry field 13 written in a specific area of the figure entry field area 22; This is an encoding unit that performs encoding, and the recognition unit 108 is configured as described above. 709 is the mark recognition result,
A computer interface unit 110 is used to edit the character recognition results and figure extraction and encoding results and transfer them to a computer, and is a computer that processes the recognition results and extracted figures.

Next, the operation shown in FIG. 1 will be explained. First, input paper 1 is input into facsimile 100. However, in this example, a G4 facsimile is used, and the input resolution is 400pp.
It was set as i. After establishing a procedure with the facsimile interface section 101, the facsimile machine 100 scans the input paper 1 and transmits a facsimile signal to the facsimile interface section 101. Facsimile interface section 10
1 detects the resolution of the image signal from the exchange of procedures with the facsimile 100, notifies the information area detection section 103, extracts the image signal from the received facsimile signal, and transfers it to the image signal memory section 102. When the image signals for one page have been transferred, the information area detection unit 103 is notified of the end of image signal storage.

The information area detecting unit 103 reads out the image signal from the image signal memory unit 102, and extracts the mark entry area 20 from the various control marks using the method shown in FIG. The character entry column area 212 goes to detect the figure entry column area 22.

When the mark entry area 20 is detected, the position on the image signal and the resolution conversion ratio (in this case, main scanning 1/2, sub-scanning 1/n4 from FIG.
Notify 04. The resolution conversion unit 104 converts the resolution of the mark entry column area 20 in the image signal memory unit 102 according to instructions from the information area detection unit 103, transfers the converted image signal to the mark recognition unit 105, and indicates the end of the conversion in the information area. The detection unit 103 is notified. The information area detection unit 103, which has received the conversion completion notification, instructs the mark recognition unit 105 to perform recognition. The mark recognition unit 105 reads the image signal at a specific position within the mark entry field area 20 and recognizes the presence or absence of an entry mark.

Then, the recognition result is transferred to the computer interface section 109, and the information area detection section 103 is notified of the end of the recognition.

When the information area detection unit 103 detects the character entry field area 21, the resolution conversion unit converts the position on the image signal and the resolution conversion ratio (1/2 in the main scan and 1/2 in the sub scan from FIG. 4). 104.

The resolution conversion unit 104 converts the resolution of the character entry area 21 in the image signal memory unit 102 according to instructions from the information area detection unit 103, transfers the converted image signal to the character recognition unit 106, and detects the information area to indicate the end of the conversion. Department 103 is notified. The information area detection unit 103, which has received the conversion completion notification, instructs the character recognition unit 106 to perform recognition. The character recognition unit 106 reads the image signal at a specific position within the character entry field area 21 and recognizes the entered character. Then, the recognition result is transferred to the computer interface unit 109, and the information area detection unit 103 indicates the end of the recognition.
Notify.

When the information area detection unit 103 detects the graphic entry field area 22, the position on the image signal and the resolution conversion ratio (in this case, from FIG. 4, main/sub-scan 1, that is, no conversion) are sent to the resolution conversion unit 104. Notice. The resolution conversion unit 104 follows instructions from the information area detection unit 103 and converts the image signal memory unit 10
2, the converted image signal is transferred to the encoding unit 107, and the information area detection unit 103 is notified of the completion of the conversion. The information area detecting unit 103, which has received the conversion completion notification, instructs the encoding unit 107 to perform extraction and encoding. Encoding N1107 is the figure entry field area 22
The image signal within a specific area within the image is encoded. Then, the encoding result is transferred to the computer interface section 109, and the information area detection section 103 is notified of the end of encoding.

The information area detection unit 103 detects the end mark 4 in the image signal, and detects all the mark entry field areas 20. Character entry field area 2
Upon receiving the notification of the completion of the recognition of 1 and the extraction and encoding of the graphic entry field area 22, the computer interface unit 109 instructs the computer 110 to send the data. Computer interface unit 109, which has received the transmission instruction, edits the recognition result and the graphic encoding information, transmits it to computer 110, and when the transmission is completed, notifies information area detection unit 103 of the completion of transmission.

During the series of processes described above, if the information area detection unit 103 cannot normally detect the control mark, the subsequent processing is immediately interrupted and the computer interface unit 109 is notified of the abnormality.

Then, the computer interface unit 109 connects the computer 110
When the transmission is completed, the information area detection unit 103 is notified of the completion of the transmission.

Note that in this embodiment, the reference position mark 2 is used for convenience of explanation, but the process can be performed even without the reference position mark 2. Further, if there is no graphic similar to the information type designation mark above the start mark 3 and below the end mark 4, and if the processing time may be slightly increased, the start mark 3 and the end mark 4 are also unnecessary.

[Effects of the Invention] As described above, the present invention includes an information area detection unit that determines the entry area and information type of specific image information and detects an area to be recognized based on the information type mark; Since it is equipped with a resolution conversion section for matching the resolution detected by the facsimile interface section and the resolution required by the recognition section, information with different recognition resolutions can be written on the same manual paper, and There is an advantage that the recognition unit does not need to have recognition parameters for each resolution. Further, even if a recognition service with a new resolution is added, only a new recognition unit is added, and there is no need to modify the conventional recognition unit, so there is an advantage that the service can be easily added.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing an example of an input form used in this invention, and FIG.
The figure shows a method of detecting information witness areas from various control marks, and FIG. 4 shows facsimile input resolutions and resolution conversion ratios for recognition types. In the figure, 1 is the input paper, 2 is the reference position mark, 3 is the start mark, 4 is the end mark, 5 and 6 are mark line designation marks, 7 and 8 are character line designation marks, and 9 and 10 are graphic line designation marks. , 11 is a mark entry field, 12 is a character entry field, 1
3 is a figure entry column, 20 is a mark entry column area, 21 is a character entry column area, 22 is a figure entry column area, 100 is a facsimile, 101 is a facsimile interface section, 102
103 is an image signal memory section, 103 is an information area detection section, 104 is a resolution conversion section, 105 is a mark recognition section, 106 is a character recognition section, 107 is an encoding section, 108 is a recognition section, 109 is a computer interface section, 110 is a It's a calculator. Figure 3 Main scanning direction

Claims (1)

[Claims] When receiving an image signal in which an information type mark that specifies the location and type of specific image information is written in a color that can be detected by a sensor, a facsimile interface unit that detects the resolution of the image on the sending side and a An image processing device equipped with a recognition unit that recognizes specific image information such as marks, characters, and figures from signals, determines the entry area and information type of the specific image information, and based on the information type mark. The present invention is characterized by comprising: an information area detection unit that detects an area to be recognized; and a resolution conversion unit that matches the resolution detected by the facsimile interface unit with the resolution required by the recognition unit. image processing device.