JP2020114028A - Image processing apparatus, image processing method, and program - Google Patents

Abstract

To provide an image processing apparatus that can accurately detect characters from a multi-valued image.SOLUTION: An image processing apparatus 1 comprises: a first detection unit 11 that detects edges of characters from a multi-valued image Im1; an edge emphasis unit 12 that emphasizes the edges of the characters included in the multi-valued image Im1 by using a result of character edge detection performed by the first detection unit 11; a second detection unit 13 that detects a character area from an edge emphasized image Im2 in which the edges of the characters are emphasized; and a separation unit 14 that separates the multi-valued image Im1 into a character image Im3 and a background image Im4 by using a result of character edge detection performed by the first detection unit 11 and a result of character area detection performed by the second detection unit 13.SELECTED DRAWING: Figure 2

Description

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

A technology called high compression PDF (Portable Document Format) is known as an image compression technology that achieves both high compression and high image quality. The high compression PDF separates a multi-valued image into an image of a character area (hereinafter, referred to as “character image”) and an image other than the character area (hereinafter, referred to as “background image”), and separates each image. This is a technique for performing suitable image processing and compression and then combining them into one image file.

In order to properly separate the character image and the background image from the multivalued image, it is necessary to accurately detect the character from the multivalued image. As a method of detecting a character from a multi-valued image, for example, a method of detecting the edge of the character by local analysis is known, focusing on the fact that the character is often expressed as a line drawing (for example, patents Reference 1). In addition, focusing on the fact that characters often form lines, a method is known in which character lines are extracted by global analysis to detect character regions included in the character lines (for example, (See Patent Document 2).

However, these methods may not be able to detect a character accurately depending on the type of character. For example, in the method of detecting the edge of a character by local analysis, the internal area of the character cannot be properly detected for a bold character, a white character, a color ground character, or the background adjacent to the character edge is conversely detected. It often happens that a part is erroneously detected as a character. In addition, the method of detecting a character region by global analysis has a problem that, for example, a character with low contrast cannot be detected appropriately, or when trying to detect a character with low contrast, it is difficult to distinguish it from the background. Therefore, there is a demand for a technique capable of detecting a character with higher accuracy from a multi-valued image.

In order to solve the above-mentioned problem, the present invention includes a first detection unit that detects an edge of a character from a multi-valued image, and a detection result of the first detection unit, and is included in the multi-valued image. Using an edge enhancement unit that enhances the edge of a character, a second detection unit that detects a character region from the multi-valued image in which the edge of the character is enhanced, and a detection result of at least the second detection unit, A separation unit that separates the multi-valued image into a character image and a background image.

According to the present invention, there is an effect that a character can be detected with high accuracy from a multi-valued image.

FIG. 1 is a block diagram showing a hardware configuration example of the image processing apparatus of the first embodiment. FIG. 2 is a block diagram showing a functional configuration example of the image processing apparatus of the first embodiment. FIG. 3 is a flowchart illustrating the flow of operations performed by the image processing apparatus according to the first embodiment. FIG. 4 is a flowchart illustrating an example of processing performed by the first detection unit. FIG. 5 is a flowchart illustrating an example of processing by the second detection unit. FIG. 6 is a diagram illustrating a specific example of the character area detection result. FIG. 7 is a flowchart illustrating an example of processing performed by the separation unit. FIG. 8 is a diagram illustrating an example of a method of determining a color ground character. FIG. 9 is a diagram illustrating another example of the method for determining a color ground character. FIG. 10 is a diagram illustrating an example of a method for determining a white character. FIG. 11 is a diagram illustrating another example of the method for determining white characters. FIG. 12 is a diagram illustrating an example of effects of the image processing apparatus according to the embodiment. FIG. 13 is a block diagram showing a functional configuration example of the image processing apparatus of the third modified example. FIG. 14 is a block diagram showing a functional configuration example of an image processing apparatus of the fourth modified example. FIG. 15 is a block diagram showing a hardware configuration example of the multifunction peripheral of the second embodiment. FIG. 16 is a block diagram showing a functional configuration example of the multi-function peripheral of the second embodiment. FIG. 17 is a diagram showing an example of an operation screen for accepting the selection of the operation mode by the user. FIG. 18 is a diagram showing an example of a relationship table that defines the correspondence relationship between operation modes and character detection operations. FIG. 19 is a diagram showing an example of an operation screen for accepting the selection of the processing speed in addition to the selection of the operation mode. FIG. 20 is a diagram showing an example of a relationship table that defines the correspondence relationship of the character detection operation with respect to the combination of the operation mode and the processing speed. FIG. 21 is a block diagram showing a functional configuration example of the multifunction peripheral of the third embodiment. FIG. 22 is a block diagram showing a configuration example of the first detection unit in the third embodiment. FIG. 23 is a diagram showing pixel comparison directions of a matrix of M×M pixels. FIG. 24 is a diagram showing an example of a block composed of N×N pixels. FIG. 25 is a diagram showing the relationship between the block of interest and the surrounding blocks. FIG. 26 is a block diagram showing a configuration example of the color determination unit. FIG. 27 is a block diagram showing a configuration example of the color processing/UCR unit. FIG. 28 is a block diagram showing a functional configuration example of the multifunction peripheral of the first modified example. FIG. 29 is a block diagram showing a configuration example of the character area correction unit. FIG. 30 is a diagram showing an example of a mask used by the mask correction circuit.

Hereinafter, specific embodiments of an image processing apparatus, an image processing method, and a program according to the present invention will be described in detail with reference to the accompanying drawings. The embodiment described below is an application example to an image processing apparatus that generates a highly compressed PDF file of a multi-valued image expressed in full color, for example. However, the present invention is not limited to this example, and can be effectively applied to various image processing apparatuses that perform processing for separating a multi-valued image into a character image and a background image.

[First Embodiment]
<Structure of image processing device>
FIG. 1 is a block diagram showing an example of the hardware configuration of the image processing apparatus 1 of this embodiment. The image processing apparatus 1 can use a computer system such as a PC (personal computer) as hardware. That is, for example, as shown in FIG. 1, the image processing apparatus 1 includes a processor such as a CPU 101, a storage device such as a RAM 102, a ROM 103, and an HDD 104, and a network I/F 105 that is a communication interface connected to a network such as a LAN. It has a configuration in which these are connected via a bus 110.

The image processing apparatus 1 according to the present embodiment acquires, for example, a multivalued image to be processed from a scanner or a host computer connected to a network via the network I/F 105. Then, the image processing apparatus 1 processes the multi-valued image to generate a highly compressed PDF file, stores the generated highly compressed PDF file in the HDD 104, or connects to the network via the network I/F 105. It is sent to the host computer. The function of generating a highly compressed PDF file from a multi-valued image is realized, for example, by the CPU 101 using the RAM 102 as a work area and executing a predetermined program stored in the ROM 103, the HDD 104, or the like.

The image processing apparatus 1 according to the present embodiment can also be realized as one function of an image forming apparatus including a scanner, such as a copying machine or a multifunction peripheral. In this case, the image forming apparatus has a computer system as shown in FIG. Then, for example, the CPU 101 inside the image forming apparatus uses the RAM 102 as a work area and executes a predetermined program stored in the ROM 103, the HDD 104, or the like to read a document by the scanner engine or acquire it via a network. The function of generating a highly compressed PDF file from the multivalued image is realized.

FIG. 2 is a block diagram showing a functional configuration example of the image processing apparatus 1 of this embodiment. The image processing apparatus 1 has, as functional components for generating a highly compressed PDF file from a multi-valued image, for example, as shown in FIG. The detection unit 13, the separation unit 14, and the file generation unit 15 are provided. In the image processing apparatus 1 of this embodiment, the multi-valued image Im1 acquired as the processing target is input to the first detection unit 11, the edge enhancement unit 12, and the separation unit 14, and the high compression corresponding to this multi-valued image Im1 is performed. The PDF file FIm is output from the file generation unit 15.

The first detection unit 11 performs processing for detecting an edge of a character in the input multi-valued image Im1 and outputs the result (hereinafter, referred to as “character edge detection result”). The process performed by the first detection unit 11 is similar to the method described in Patent Document 1, for example, and the edge (line drawing) of a character included in the multi-valued image Im1 is detected by local analysis on the multi-valued image Im1. This is the process of detecting. This processing has the characteristics that it can efficiently detect small letters and low-contrast characters, but cannot detect the internal area of bold characters, and is prone to erroneous detection in the detection of ground-colored characters and white characters. Therefore, when the multi-valued image Im1 to be processed includes a bold character, a color ground character, a white character, etc., the character edge detection result output by the first detection unit 11 detects the internal area of the bold character. Leakage, erroneous detection around the edge of the color ground character or white character, etc. may occur.

Here, the character edge detection result is, for example, coordinate data representing the coordinate position in the multi-valued image Im1 of the pixel group detected as the edge of the character by the first detection unit 11. The character edge detection result is input to the edge enhancing unit 12 and the separating unit 14. The details of the processing by the first detection unit 11 will be described later.

The edge emphasizing unit 12 uses the character edge detection result input from the first detecting unit 11 to perform the process of emphasizing the edge of the character included in the multivalued image Im1 to be processed, and the edge of the character is emphasized. A multi-valued image (hereinafter referred to as “edge-enhanced image”) Im2 is output. This edge-enhanced image Im2 is input to the second detection unit 13.

The second detection unit 13 performs a process of detecting a character area on the edge-emphasized image Im2 input from the edge emphasis unit 12, and outputs the result (hereinafter, referred to as “character area detection result”). In the processing by the second detection unit 13, a character line is extracted from the edge-emphasized image Im2 by global analysis, and the character region included in the character line is detected, as in the method described in Patent Document 2, for example. Processing. This process has a characteristic that it is possible to accurately detect a bold character, a ground character, a white character, and the like, but it is difficult to correctly detect a gray character having a low contrast. Therefore, when the multi-valued image Im1 to be processed includes gray characters or the like, the character area detection result output by the second detection unit 13 may include omission of gray characters or the like.

Here, the character area detection result is, for example, coordinate data representing the coordinate position in the multi-valued image Im1 of the pixel group detected as the character area by the second detection unit 13. The character area detection result is input to the separation unit 14. The details of the processing by the second detection unit 13 will be described later.

The separation unit 14 uses the character edge detection result input from the first detection unit 11 and the character region detection result input from the second detection unit 13 to convert the multi-valued image Im1 into the character image Im3 and the background image Im4. And separate. Specifically, the separating unit 14 detects the area of the color ground character and the area of the white character in the multi-valued image Im1 by using the character area detection result and the multi-valued image Im1, and adds the color to the character edge detection result. If there is a part detected as an edge of a ground character or an edge of a white character, remove it. Then, the separation unit 14 integrates (merges) the character edge detection result obtained by removing the edge of the color ground character and the edge of the white character and the character region detection result by the OR operation, and the multi-value indicated by the integrated detection result. A character image Im3 is obtained by extracting the image area in the image Im1. Further, the separation unit 14 sets the difference between the multi-valued image Im1 and the character image Im3 as the background image Im4. The character image Im3 and the background image Im4 are input to the file generation unit 15.

In the present embodiment, when the multi-valued image Im1 includes both the area of the color ground character and the area of the white character, the separating unit 14 determines the area of both the area of the color ground character and the area of the white character. Is detected and both the edge of the ground character and the edge of the white character are removed from the character edge detection result, but the present invention is not limited to this. The separating unit 14 may be configured to detect only the area of the color ground character included in the multi-valued image Im1 and remove only the edge of the color ground character from the character edge detection result, or the multi-valued image Im1. It is also possible to detect only the area of the white character included in the above and remove only the edge of the white character from the character edge detection result.

The file generation unit 15 compresses the character image Im3 and the background image Im4 input from the separation unit 14 by different methods and then integrates them to generate a high-compression PDF file FIm corresponding to the multi-valued image Im1. Specifically, for example, the file generation unit 15 performs, on the character image Im3, a color reduction process that limits the number of colors, and then compresses the image by an encoding method such as MMR. On the other hand, with respect to the background image Im4, the amount of data is reduced by downsampling and then compressed by an encoding method such as JPEG. Then, the compressed character image Im3 and the compressed background image Im4 are integrated on one image file in PDF format, for example, to generate a highly compressed PDF file FIm corresponding to the multi-valued image Im1.

The method of compressing the character image Im3 and the background image Im4 is not limited to the above example, and compression may be performed by a method suitable for each of the character image Im3 and the background image Im4. Further, the format of the image file that integrates the compressed character image Im3 and the compressed background image Im4 is not limited to the PDF format. For example, the JPM format or the like is used to overlap the background and the foreground to form one image. Various formats can be used.

The high compression PDF file FIm generated by the file generation unit 15 is, for example, accumulated in the HDD 104 or transmitted to the host computer or the like connected to the network via the network I/F 105, as described above.

<Operation of image processing device>
Next, an outline of the operation of the image processing apparatus 1 of the present embodiment configured as described above will be described. FIG. 3 is a flowchart illustrating the flow of operations performed by the image processing apparatus 1 according to this embodiment.

When the operation of the image processing apparatus 1 of the present embodiment is started, first, in step S1, the multivalued image Im1 to be processed is acquired. The multi-valued image Im1 is input to the first detection unit 11, the edge enhancement unit 12, and the separation unit 14.

Next, in step S2, the process by the first detection unit 11 is performed on the multi-valued image Im1 acquired in step S1. Then, the character edge detection result output by the first detection unit 11 is input to the edge emphasis unit 12 and the separation unit 14.

Next, in step S3, the edge emphasis unit 12 uses the character edge detection result of step S2 to emphasize the edge of the character included in the multi-valued image Im1 acquired in step S1. Im2 is generated. This edge-enhanced image Im2 is input to the second detection unit 13.

Next, in step S4, the second detection unit 13 performs the process on the edge-enhanced image Im2 generated in step S3. Then, the character area detection result output by the second detection unit 13 is input to the separation unit 14.

Next, in step S5, the separation unit 14 uses the character edge detection result of step S2 and the character area detection result of step S4 to convert the multi-valued image Im1 acquired in step S1 into a character image Im3 and a background image Im4. The process of separating into. The character image Im3 and the background image Im4 are input to the file generation unit 15.

Next, in step S6, the file generation unit 15 performs appropriate image processing and compression on the character image Im3 and the background image Im4 generated in step S5, and then merges them into one image file. Is performed, and the high compression PDF file FIm corresponding to the multi-valued image Im1 acquired in step S1 is generated.

Finally, in step S7, the highly compressed PDF file FIm generated in step S6 is output from the file generation unit 15 and stored in, for example, the HDD 104, or a host connected to the network via the network I/F 105. Sent to a computer, etc.

<Specific Example of Processing by First Detecting Unit>
Next, a specific example of the processing by the first detection unit 11 will be described with reference to FIG. FIG. 4 is a flowchart illustrating an example of processing performed by the first detection unit 11.

As described above, the processing performed by the first detection unit 11 is processing for detecting the edge of a character included in the multi-valued image Im1 by local analysis, and, for example, using the technique described in Patent Document 1, This can be realized by detecting the line drawing included in the multi-valued image Im1 as the edge of the character.

The first detection unit 11 first performs MTF correction by filtering on the multivalued image Im1 to be processed (step S101). The MTF correction is a pre-process that is performed to improve the accuracy of the next ternarization. As the filter for MTF correction, for example, the filter described in Patent Document 1 can be used.

Next, the first detection unit 11 performs ternarization on the MTF-corrected multi-valued image Im1 (step S102) and divides the multi-valued image Im1 into black pixels, white pixels, and gray pixels.

Next, the first detection unit 11 extracts a black line drawing or a white line drawing from the multi-valued image Im1 by pattern matching of black connected pixels or white connected pixels, as described in Patent Document 1, for example, The edge of the character included in the multi-valued image Im1 is detected (step S103). At this time, by appropriately setting the threshold value in the ternarization in step S102, even when the multi-valued image Im1 includes a gray character, the edge of the gray character can be appropriately detected.

Then, the first detection unit 11 outputs the processing result of step S103 as the character edge detection result (step S104), and ends the series of processing.

<Specific Example of Processing by Second Detection Unit>
Next, a specific example of the processing by the second detection unit 13 will be described with reference to FIG. FIG. 5 is a flowchart illustrating an example of processing performed by the second detection unit 13.

As described above, the process performed by the second detection unit 13 is a process of detecting a character region from an image by global analysis, and for example, using the technique described in Patent Document 2, a character is detected from the edge-enhanced image Im2. It can be realized by extracting a line and detecting a character area from the character line. Here, in the image processing device 1 of the present embodiment, the multi-valued image Im1 to be processed is not processed by the second detection unit 13, but is a character that is the processing result of the first detection unit 11. The second detection unit 13 performs processing on the edge-enhanced image Im2 generated by the edge enhancement unit 12 using the edge detection result. Therefore, the detection accuracy of the character area is higher than that in the case of detecting the character area as it is from the multi-valued image Im1.

The second detection unit 13 first binarizes the edge-emphasized image Im2 generated by the edge enhancement unit 12 (step S201). The binarization here is a process for extracting a character line having a lower brightness (for example, black) than the background, and an appropriate threshold value that can distinguish the low brightness character line from the background is set. Also, dynamic threshold binarization may be used to increase the accuracy of separation from the background.

Next, as described in, for example, Patent Document 2, the second detection unit 13 determines, from the binarized image obtained in step S201, a run of black pixels arranged in the horizontal direction and a black pixel arranged in the vertical direction. The runs are connected to obtain a connected component (step S202).

Next, the second detection unit 13 removes, from the connected components acquired in step S202, connected components that are estimated to be small-sized characters (hereinafter referred to as "lowercase letters") (step S203). This process is a process for reducing the calculation amount in the extraction of the next character line. That is, extraction of a character line is a process of integrating the acquired connected components as a character line according to various conditions. However, when a large number of connected components estimated to be in lowercase are included, the combination of the connected components is combined. The number of patterns will become huge and the amount of calculation will increase significantly. On the other hand, a low-luminance (eg, black) lowercase letter included in the multi-valued image Im1 is likely to be detected by the first detection unit 11 and output as a character edge detection result, and detected by the second detection unit 13. Even if it is excluded from the target, it is assumed that there will be no omission in detection. Therefore, in the present embodiment, the amount of calculation is reduced by removing the connected component that is estimated to be a low-luminance lowercase letter and then extracting the next character line. It should be noted that whether or not the letter is a low-luminance lowercase letter is determined by, for example, the size of the circumscribed rectangle of the connected component acquired from the binarized image in step S202 being equal to or smaller than a predetermined value, and the circumscribed rectangle near the connected component. May be performed according to the criterion that there is no connected component that exceeds a predetermined value.

Next, the second detection unit 13 integrates the connected components acquired in step S202 and not removed in step S203 by, for example, the method described in Patent Document 2 to form a character line having a lower luminance than the background. Extract (step S204).

Next, the second detection unit 13 re-binarizes the edge emphasized image Im2 generated by the edge emphasis unit 12 (step S205). The binarization here is a process for extracting a character line having a higher brightness (for example, white) than the background, and an appropriate threshold value that can distinguish the high brightness character line from the background is set. Also, dynamic threshold binarization may be used to increase the accuracy of separation from the background.

Next, as described in, for example, Patent Document 2, the second detection unit 13 uses the run of white pixels arranged in the horizontal direction and the white pixels arranged in the vertical direction from the binarized image obtained in step S205. The runs are connected to obtain a connected component (step S206).

Next, the second detection unit 13 integrates the connected components acquired in step S206 by, for example, the method described in Patent Document 2 to extract a character line having a higher brightness than the background (step S207).

Next, the second detection unit 13 determines whether or not there is a character line whose position in the edge-enhanced image Im2 overlaps with the character line extracted in step S204 and the character line extracted in step S207 (step S208). .. If there are overlapping character lines (step S208: Yes), the sizes of the circumscribing rectangles of these character lines are compared, and the character line having the smaller circumscribing rectangle size is deleted (step S209).

Next, the 2nd detection part 13 detects the character contained in the character line extracted from the edge emphasis image Im2 by the above process as a character area (step S210). Then, the second detection unit 13 outputs the processing result of step S210 as the character area detection result (step S211), and ends the series of processing.

FIG. 6 is a diagram illustrating a specific example of the character area detection result obtained by the processing by the second detection unit 13. In this FIG. 6, an example in which a character area “aiueo” of black characters is detected from the edge emphasized image Im2 is shown.

It is assumed that a binarized image such as 601 in FIG. 6 is obtained by the processing in step S201 in FIG. When the process of step S202 in FIG. 5 is performed on this binarized image, the connected component indicated by 602 in FIG. 6 is acquired. Then, by the processing of step S203 in FIG. 5, the lowercase connected components are removed, and the connected components such as 603 in FIG. 6 remain. By performing the process of step S204 of FIG. 5 on this connected component, a character line is extracted as indicated by 604 of FIG. Then, by the process of step S210 of FIG. 5, the character area of “aiueo” included in the character line is detected as in 605 of FIG. 6, and is output as the character area detection result.

<Specific example of processing by the separation unit>
Next, a specific example of the processing performed by the separation unit 14 will be described with reference to FIG. 7. FIG. 7 is a flowchart illustrating an example of processing performed by the separating unit 14.

As described above, the processing by the separation unit 14 uses the character edge detection result by the first detection unit 11 and the character area detection result by the second detection unit 13 to convert the multi-valued image Im1 into the character image Im3. This is a process of separating the background image Im4. Here, in the image processing device 1 of the present embodiment, the separating unit 14 first detects the area of the color ground character and the area of the white character in the multi-valued image Im1, and detects the edge of the color ground character in the character edge detection result. Alternatively, if there is a portion detected as the edge of the white character, it is removed. As a result, in the processing of the first detection unit 11, it is possible to remove the edges of the color ground characters and white characters that are likely to be erroneously detected, and improve the reliability of the character edge detection result. Then, the character edge detection result and the character area detection result with improved reliability are integrated, and the multi-valued image Im1 is separated into the character image Im3 and the background image Im4 by using the integrated detection result, thereby separating the images. The accuracy can be increased.

First, the separation unit 14 determines whether the multi-valued image Im1 includes a color ground character by using the character area detection result input from the second detection unit 13 and the multi-valued image Im1 (step S301). ).

FIG. 8 is a diagram illustrating an example of a method of determining a color ground character. FIG. 8 shows an example in which the black character “aiueo” exists on the yellow ground. Reference numeral 801 in FIG. 8 indicates a multi-valued image Im1 of a circumscribed rectangle of a character line including “aiueo” extracted by the second detection unit 13. 8 represents the position, and 802 in FIG. 8 represents the character area “aiueo” detected by the second detection unit 13.

The separating unit 14 removes the character area “Aiueo” shown by 802 in FIG. 8 from the circumscribed rectangle on the multi-valued image Im1 shown by 801 in FIG. Can be extracted. Then, it is possible to determine whether or not the ground value character is included in the multi-valued image Im1 by determining whether or not the color of the base region is a chromatic color. In the example of FIG. 8, since the color of the background area is yellow, it is determined that the multi-valued image Im1 includes color ground characters.

FIG. 9 is a diagram for explaining another example of the method for determining a color ground character, and shows a method for determining whether a character close to a rectangle is a color ground character. In this FIG. 9, an example in which a black long note “−” exists on the yellow ground is shown, and 901 in FIG. 9 is a multi-valued circumscribed rectangle of a character line including “−” extracted by the second detection unit 13. The position in the image Im1 is represented, and 902 in FIG. 9 represents the character area “−” detected by the second detection unit 13.

As in the example of FIG. 9, when the character area detected by the second detection unit 13 is close to a rectangle, the background area of the character area cannot be appropriately extracted by the method shown in FIG. Therefore, when the character region detected by the second detection unit 13 is close to a rectangle, the separation unit 14 determines the peripheral region close to the character region as a determination target, as shown in 903 of FIG. By determining whether or not the color is a chromatic color, it is determined whether or not the multi-valued image Im1 includes color ground characters. In the example of FIG. 9, since the color of the peripheral area is yellow, it is determined that the multi-valued image Im1 includes color ground characters.

When the separation unit 14 determines in step S301 that the multi-valued image Im1 includes color ground characters (step S301: Yes), the character edge detection result by the first detection unit 11 is collated with the multi-valued image Im1. Then, if the character edge detection result includes a portion in which the edge of the color ground character is detected, the edge of the color ground character is removed (step S302). On the other hand, if the multi-valued image Im1 does not include the color ground characters (step S301: No), the process proceeds to step S303.

Next, the separating unit 14 uses the character region detection result input from the second detecting unit 13 and the multi-valued image Im1 to determine whether the multi-valued image Im1 includes white characters (step S303). ).

FIG. 10 is a diagram illustrating an example of a method for determining a white character. In this FIG. 10, an example in which a white character “AIUEO” exists on a black ground is shown, and 1001 in FIG. 10 indicates the position in the multi-valued image Im1 of the character area “AIUEO” detected by the second detection unit 13. It represents.

The separating unit 14 extracts the internal area of the character from the character area “aiueo” on the multi-valued image Im1 shown by 1001 in FIG. 10 as shown by 1002 in FIG. Then, by determining whether or not the color of the internal area of the extracted character is white, it is possible to determine whether or not the multi-valued image Im1 includes a white character. Here, the target of the internal area of the character is that if the edge of the character is included in the target of the character color determination, the character color cannot be correctly determined when the background color is mixed in the edge of the character. This is because. In the example of FIG. 10, since the color of the internal area of the character is white, it is determined that the multi-valued image Im1 includes white characters.

FIG. 11 is a diagram for explaining another example of the method for determining a white character, and shows the determination method when the character to be determined is a thin character. FIG. 11 shows an example in which a thin white character “aiueo” exists on a black ground, and 1101 in FIG. 11 indicates that many circumscribed rectangles of a character line including “aiueo” extracted by the second detection unit 13 are included. The position in the value image Im1 is represented, and 1102 in FIG. 11 represents the character area “aiueo” detected by the second detection unit 13.

As in the example of FIG. 11, when the character region detected by the second detection unit 13 is a thin character, the internal region of the character cannot be properly extracted as in the example shown in FIG. Therefore, when the character region detected by the second detection unit 13 is a thin character, the separation unit 14 determines from the circumscribed rectangle on the multi-valued image Im1 1101 in FIG. 11 to the character region “1102 in FIG. By removing "aiueo", the background area of the character area "aiueo" is extracted as indicated by 1103 in FIG. Then, by using the color of the background area and the color of the character area, it is determined whether or not the color of the character area is a color similar to the color of the background area and has a higher brightness than the background area. It is determined whether Im1 includes a white character. Here, one of the conditions is that the color of the character area is similar to the color of the background area. Even if the background color is mixed at the edge of the character, it can be determined whether the character color is white. To do so. In the example of FIG. 11, since the color of the character area (white) is similar to the color of the background area (black) and has a higher brightness than the background area, it is determined that the multi-valued image Im1 contains white characters. Will be done.

When it is determined in step S303 that the multi-valued image Im1 includes white characters (step S303: Yes), the separation unit 14 collates the character edge detection result by the first detection unit 11 with the multi-valued image Im1. If the character edge detection result includes a portion in which the edge of the white character is detected, the edge of the white character is removed (step S304). On the other hand, if the multi-valued image Im1 does not include white characters (step S303: No), the process proceeds to step S305.

Next, the separation unit 14 detects the character edge detection result obtained by removing the edge of the ground character and the edge of the white character from the character edge detection result obtained by the first detection unit 11, and the character area detection result obtained by the second detection unit 13. The two are integrated by OR operation with. Then, the separation unit 14 generates the character image Im3 by extracting the image area indicated by the integrated detection result from the multi-valued image Im1 (step S305). Further, the separation unit 14 generates a background image Im4 by using the difference between the character image Im3 generated in step S305 and the multi-valued image Im1 (step S306).

Then, the separation unit 14 outputs the character image Im3 and the background image Im4 generated in steps S305 and S306 (step S307), and ends the series of processes.

<Effects of the embodiment>
As described above in detail with reference to specific examples, the image processing apparatus 1 according to the present embodiment uses the first detection unit 11 that detects the edge of a character by the local analysis and the global analysis. A second detection unit 13 for detecting a character region is provided, and the character edge detection result by the first detection unit 11 and the character region detection result by the second detection unit 13 are used in combination to finally The character area included in the multi-valued image Im1 to be processed is detected. Therefore, according to the image processing apparatus 1 of the present embodiment, even if the multi-valued image Im1 to be processed includes various types of characters, it is possible to detect a character from the multi-valued image Im1 with high accuracy. it can.

Particularly, in the image processing device 1 of the present embodiment, the edge enhancement unit 12 enhances the edge enhancement by enhancing the edges of the characters included in the multi-valued image Im1 by using the character edge detection result by the first detection unit 11. The image Im2 is generated. Then, the second detection unit 13 detects the character area from the edge emphasized image Im2 generated by the edge emphasis unit 12. Therefore, the detection accuracy of the character area by the second detection unit 13 can be improved.

Further, in the image processing apparatus 1 according to the present embodiment, the separation unit 14 integrates the character edge detection result by the first detection unit 11 and the character area detection result by the second detection unit 13, and outputs the integrated detection result. The multi-valued image Im1 to be processed is separated into the character image Im3 and the background image Im4. Therefore, the multi-valued image Im1 can be accurately separated into the character image Im3 and the background image Im4.

Further, in the image processing apparatus 1 of the present embodiment, the separation unit 14 removes the edges of the color ground characters and the white characters from the character edge detection result by the first detection unit 11 and then the characters by the second detection unit 13. It is designed to be integrated with the area detection result. Therefore, even when the character edge detection result by the first detection unit 11 includes an erroneous detection near the edge of a color ground character or a white character, the accuracy of the character image Im3 is affected by the erroneous detection. The reduced inconvenience can be effectively suppressed.

In addition, in the image processing apparatus 1 of the present embodiment, the second detection unit 13 removes the connected component estimated to be lowercase letters from the connected components acquired from the edge-enhanced image Im2, and then performs integration into a character line. I'm trying to do it. Therefore, it is possible to reduce the calculation amount required for extracting the character line and reduce the processing load of the second detection unit 13.

FIG. 12 is a diagram for explaining an example of the effect of the image processing apparatus 1 of the present embodiment, and shows specific examples of the character image Im3 and the background image Im4 generated from the multivalued image Im1 to be processed. Reference numeral 1201 in FIG. 12 shows a specific example of the character edge detection result obtained by the processing of the first detection unit 11, and 1202 of FIG. 12 shows a specific example of the character area detection result obtained by the processing of the second detection unit 13. An example is shown.

As shown in the example of FIG. 12, when the multi-valued image Im1 to be processed includes a bold character, the first detection unit 11 cannot properly detect the internal area of the bold character, so the character edge detection result 1201 Is missing in the internal area of bold characters. In the case where the multi-valued image Im1 to be processed includes a color ground character and a white character (FIG. 12 shows an example in which a color ground character and a white character are included in the multi-valued image Im1), Since the detection unit 11 of No. 1 is likely to make an erroneous detection when detecting a color ground character or a white character, the character edge detection result 1201 includes an erroneous detection near the edge of a white character (color ground character). When the multi-valued image Im1 to be processed includes gray characters, the second detection unit 13 cannot properly detect low-contrast characters, and therefore the character region detection result 1202 shows the gray character regions. Is not included.

However, in the image processing apparatus 1 of the present embodiment, the character edge detection result by the first detection unit 11 is removed from the character edge detection result 1201 by the first detection unit 11, and then the character area detection result by the second detection unit 13 is removed. The character image Im3 is generated by integrating with 1202 and extracting the image area indicated by the integrated detection result from the multi-valued image Im1. Then, the difference between the character image Im3 and the multi-valued image Im1 is generated as the background image Im4. Therefore, as shown in the example of FIG. 12, even when the multi-valued image Im1 to be processed includes various types of characters such as bold characters, ground characters, white characters, and gray characters, A character can be detected from the multi-valued image Im1 with high accuracy, and the multi-valued image Im1 can be accurately separated into the character image Im3 and the background image Im4.

<First Modification>
In the first embodiment described above, when the second detection unit 13 extracts character lines, the connected components that are estimated to be lowercase letters are removed and the remaining connected components are integrated as character lines. I am trying. However, all the acquired connected components may be integrated into a character line without removing such connected components estimated to be lowercase. In this case, the processing load on the second detection unit 13 is increased as compared with the case of removing the connected component estimated to be the lowercase character, but the lowercase character included in the multi-valued image Im1 can be detected more accurately. ..

<Second Modification>
Further, in the above-described first embodiment, the separating unit 14 removes the edge portions of the ground character and the white character from the character edge detection result by the first detecting unit 11, and then the character by the second detecting unit 13. The area detection result 1202 is integrated. However, the character edge detection result and the character area detection result may be integrated without removing the edge portion of the color ground character or the white character. In this case, there is a possibility that the integrated detection result may be affected by the erroneous detection by the first detection unit 11, but the processing load on the separation unit 14 can be reduced and the processing time can be shortened.

<Third Modification>
Further, in the above-described first embodiment, the character edge detection result by the first detection unit 11 is used for both the generation of the edge emphasized image Im2 by the edge emphasis unit 12 and the generation of the character image Im3 by the separation unit 14. I am trying to use it. However, the character edge detection result by the first detection unit 11 may be used only for generating the edge-enhanced image Im2 and not for generating the character image Im3.

FIG. 13 is a block diagram showing a functional configuration example of the image processing apparatus 1 ′ of this modification. As shown in FIG. 13, in the image processing apparatus 1 ′ of this modification, the character edge detection result by the first detection unit 11 is input only to the edge emphasis unit 12 and not to the separation unit 14. Therefore, the separation unit 14 does not integrate the character edge detection result by the first detection unit 11 and the character area detection result by the second detection unit 13, and the character area detection result by the second detection unit 13 indicates. The character image Im3 is generated by extracting the image area from the multi-valued image Im1.

Even if the separation unit 14 generates the character image Im3 using only the character area detection result by the second detection unit 13 as in the present modification, the second detection unit 13 does not detect the character area. , The edge emphasized image Im2 generated by the edge emphasizing unit 12 using the character edge detection result by the first detecting unit 11 is targeted, so that the character area can be accurately detected, and the multi-valued image Im1 can be detected. The image Im3 and the background image Im4 can be accurately separated.

<Fourth Modification>
Further, in the above-described first embodiment, the character edge detection result by the first detection unit 11 is used for both the generation of the edge emphasized image Im2 by the edge emphasis unit 12 and the generation of the character image Im3 by the separation unit 14. I am trying to use it. However, the character edge detection result by the first detection unit 11 may be used only for generating the character image Im3, and the edge emphasized image Im2 may not be generated.

FIG. 14 is a block diagram showing a functional configuration example of the image processing apparatus 1 ″ of this modification. As shown in FIG. 14, in the image processing apparatus 1 ″ of this modification, the edge enhancement unit 12 is not provided, and the character edge detection result by the first detection unit 11 is input only to the separation unit 14. .. Therefore, the second detection unit 13 does not perform the process of detecting the character region for the edge-enhanced image Im2 as in the above-described embodiment, but performs the process of detecting the character region for the multi-valued image Im1. To do.

Even if the second detection unit 13 is configured to detect the character region for the multi-valued image Im1 as in this modification, the character edge detection result by the first detection unit 11 in the separation unit 14 is detected. And the character area detection result of the second detection unit 13 are integrated, and the character area included in the multi-valued image Im1 is finally specified using the integrated detection result. Im3 and the background image Im4 can be accurately separated.

<Supplementary explanation>
Functional components of the image processing apparatus 1 described in the above-described first embodiment and modification (first detection unit 11, edge enhancement unit 12, second detection unit 13, separation unit 14, and file generation unit 15). ) Can be realized by the cooperation of the hardware and software (program) shown in FIG. 1, for example, as described above. In this case, the program is recorded in a computer-readable recording medium such as a CD-ROM, a flexible disk (FD), a CD-R, or a DVD as a file in a format installable or executable in the image processing apparatus 1. Will be provided. Further, the program may be stored in a computer connected to a network such as the Internet and provided by being downloaded to the image processing apparatus 1 via the network. Furthermore, the above program may be configured to be provided or distributed via a network such as the Internet. Further, the above program may be provided by being previously incorporated in the ROM 103, the HDD 104, or the like in the image processing apparatus 1, for example.

In addition, some or all of the functional components of the image processing apparatus 1 described in the above-described first embodiment and modified example are, for example, ASIC (Application Specific Integrated Circuit) and FPGA (Field-Programmable Gate Array). It can also be realized by using dedicated hardware such as.

Further, in the above-described first embodiment and modified example, an example in which the image processing device 1 is realized as a single device is assumed, but the functional components of the image processing device 1 may be physically separated into a plurality of devices. The image processing apparatus 1 may be provided in a distributed manner and the operation of the image processing apparatus 1 may be realized by cooperation of these plural apparatuses.

[Second Embodiment]
Next, a second embodiment will be described. The present embodiment is an example in which whether or not to execute the process by the second detection unit 13 that detects the character region by the global analysis can be switched according to the operation mode selected by the user. In the present embodiment, an example in which the above-described function as the image processing apparatus is realized in a multifunction peripheral (MFP) is given. In this multi-function peripheral, the first detection unit 11 that detects the edge of the character by local analysis is realized by hardware implementation such as ASIC, and the second detection unit 13 that detects the character region by global analysis is used. It shall be realized by software implementation.

In the above-described first embodiment, by combining the processing by the first detection unit 11 and the processing by the second detection unit 13, a character is detected with high accuracy from the multivalued image Im1 to be processed, and the multivalued image is obtained. Separation performance is improved when Im1 is separated into the character image Im3 and the background image Im4. Although this method provides a high degree of satisfaction to users who place importance on character image quality, it takes a relatively long time to perform processing, which may cause dissatisfaction to users who place importance on productivity. In particular, it is difficult to mount the second detection unit 13 that detects the character region by the global analysis in hardware, and it is generally realized by software. For this reason, the processing of the second detection unit 13 requires a relatively long time, which causes a user who attaches importance to productivity to be dissatisfied.

Therefore, in the present embodiment, operation modes selectable by the user, such as “standard mode” and “character priority mode” (or “picture priority mode” and “character priority mode”), are prepared. Then, according to the operation mode selected by the user, whether to execute the process by the second detection unit 13, that is, whether to activate or deactivate the second detection unit 13 is switched. ing. In the above example, when the “standard mode” is selected by the user, the second detection unit 13 is inactive, and when the “character priority mode” is selected by the user, the second detection unit 13 is active. It is said that

Since the first detection unit 11 is implemented by hardware, high speed processing is possible. Therefore, regardless of the operation mode selected by the user, the processing by the first detection unit 11 is always executed. In the above example, when the "standard mode" is selected by the user, the first detection unit 11 is active, the second detection unit 13 is inactive, and the "character priority mode" is selected by the user. Both the first detection unit 11 and the second detection unit 13 are activated.

FIG. 15 is a block diagram showing an example of the hardware configuration of the multifunction machine 200 of this embodiment. As shown in, for example, FIG. 15, the multifunction device 200 of the present embodiment includes a controller 210, an operation panel 220, an FCU (Facsimile Control Unit) 230, a USB (Universal Serial Bus) device 240, and an MLB (Media Link Board). ) 250, a scanner engine 260, and a plotter engine 270.

The operation panel 220 is a user interface that allows the user of the multifunction peripheral 200 to input various settings and display various information presented to the user. The above-described operation mode selection by the user is performed using the operation panel 220, for example.

The FCU 230 is a control unit that controls the facsimile function of the multi function device 200. The USB device 240 is a device connected to the multifunction device 200 by USB. The MLB 250 is a conversion board that converts the format of image data. The scanner engine 260 is an engine for reading a document, and the plotter engine 270 is an engine for printing. In the present embodiment, it is assumed that the multi-valued image Im1 to be processed is acquired by reading the document with the scanner engine 260.

The controller 210 is a control device that controls the operation of the multifunction peripheral 200. As shown in FIG. 15, the controller 210 includes a CPU 211, a system memory 212, a HDD (Hard Disk Drive) 213, a PHY 214, and an ASIC 215. The operation panel 220 is connected to the ASIC 215 of the controller 210. Further, the FCU 230, the USB device 240, the MLB 250, the scanner engine 260, and the plotter engine 270 are connected to the ASIC 215 of the controller 210 via the data transfer bus 280.

In the multifunction machine 200 of this embodiment, the function as the image processing apparatus described above is realized mainly by the controller 210. That is, of the functional components of the image processing apparatus 1 of the first embodiment shown in FIG. 2, the first detection unit 11 and the edge enhancement unit 12 are realized by, for example, the ASIC 215 of the controller 210. The second detection unit 13, the separation unit 14, and the file generation unit 15 are realized, for example, by the CPU 211 of the controller 210 executing a predetermined program (software) using the system memory 212.

FIG. 16 is a block diagram showing an example of the functional configuration of the multifunction machine 200 of this embodiment. The multifunction device 200 of the present embodiment includes, for example, as shown in FIG. 16, a first processing unit 310, a second processing unit 320, and a switching unit 330.

The first processing unit 310 is a functional module implemented by the ASIC 215 of the controller 210, and includes the first detection unit 11 and the edge enhancement unit 12. The first detection unit 11 and the edge enhancement unit 12 are the same as in the above-described first embodiment. That is, the first detection unit 11 executes, for example, the process shown in FIG. 4 on the multi-valued image Im1 to be processed (the image read by the scanner engine 260 in the present embodiment), so that the multi-valued image Im1. To detect the edges of the characters contained in. Further, the edge emphasizing unit 12 performs a process of emphasizing the edges of the characters included in the multi-valued image Im1 by using the character edge detection result of the first detecting unit 11 to generate the edge emphasizing image Im2.

The second processing unit 320 is a functional module realized by the CPU 211 of the controller 210 executing a predetermined program (software) using the system memory 212, and includes the second detection unit 13, the separation unit 14, and the The file generation unit 15 is included. The second detection unit 13, the separation unit 14, and the file generation unit 15 are the same as those in the second embodiment described above. That is, the second detection unit 13 detects the character region included in the multi-valued image Im1 to be processed by executing, for example, the process shown in FIG. 5 on the edge-enhanced image Im2. Further, the separation unit 14 uses the character edge detection result by the first detection unit 11 and the character area detection result by the second detection unit 13 to execute the process shown in FIG. The value image Im1 is separated into a character image Im3 and a background image Im4. In addition, the file generation unit 15 compresses the character image Im3 and the background image Im4 separated from the multi-valued image Im1 by the separation unit 14 by different methods and then integrates them into a high compression PDF file corresponding to the multi-valued image Im1. Generate FIm.

However, in the present embodiment, the switching unit 330 switches whether to execute the process by the second detection unit 13, that is, whether to activate or deactivate the second detection unit 13. Then, the separation unit 14 performs the same process as in the first embodiment when the second detection unit 13 is active, but when the second detection unit 13 is inactive, the separation unit 14 uses the first detection unit 11. The multi-valued image Im1 is separated into a character image Im3 and a background image Im4 based on only the character edge detection result.

Data is transferred from the first processing unit 310 to the second processing unit 320 via the HDD 213 (storage unit). That is, the data to be passed from the first processing unit 310 to the second processing unit 320 (the multivalued image Im1 to be processed, the edge emphasized image Im2 generated by the edge emphasis unit 12 and the character edge by the first detection unit 11). The detection result) is temporarily stored in the HDD 213. Then, the second detection unit 13 and the separation unit 14 of the second processing unit 320 read necessary data from the HDD 213 as needed and perform the above-described processing. At this time, in order to effectively use the capacity of the HDD 213, an arbitrary compression process is performed on the data stored in the HDD 213, the compressed data is read from the HDD 213, and the decryption process is performed.

As described above, by passing the data from the first processing unit 310 to the second processing unit 320 via the HDD 213, the HDD 213 is transferred to the processing speed of the first processing unit 310 and the second processing unit 320. It can function as a buffer that absorbs the speed difference from the processing speed of 320. That is, the processing speed of the second processing unit 320 realized by software implementation becomes longer than the processing speed of the first processing unit 310 realized by hardware implementation using the ASIC 215. Therefore, if there is no buffer that absorbs the speed difference between the processing speed of the first processing unit 310 and the processing speed of the second processing unit 320, for example, a document having a large number of pages is read by the scanner engine 260 and highly compressed. When the job of generating the PDF file FIm is executed, the processing of the first processing unit 310 on the subsequent page cannot be executed until the processing of the second processing unit 320 on the preceding page is completed. As a result, it takes time to read the document by the scanner engine 260, which causes the user to wait for a long time at the installation location of the multifunction machine 200. On the other hand, by adopting a configuration in which the HDD 213 absorbs the speed difference between the processing speed of the first processing unit 310 and the processing speed of the second processing unit 320, the reading of the document by the scanner engine 260 can be performed in a short time. By doing so, the user can be released from the installation place of the multi function device 200 at an early stage.

The switching unit 330 determines whether to execute the process by the second detection unit 13 according to the operation mode selected by the user, that is, whether to activate or deactivate the second detection unit 13. Switch. The operation mode is selected by the user through, for example, an operation screen displayed on the operation panel 220.

FIG. 17 is a diagram showing an example of an operation screen for accepting the selection of the operation mode by the user. The operation screen 170 shown in FIG. 17 is a screen displayed on the operation panel 220 when a scanner job is executed. In the operation mode selection area 171, a “standard mode” button 172 and a “character priority mode” that can be touch-operated by the user are displayed. Button 173 is provided. When the user touches the “standard mode” button 172 on the operation screen 170, a mode selection signal designating the “standard mode” as the operation mode is transmitted from the operation panel 220 to the switching unit 330. On the other hand, when the user touches the “character priority mode” button 173 on the operation screen 170, a mode selection signal designating the “character priority mode” as the operation mode is sent from the operation panel 220 to the switching unit 330.

The switching unit 330 determines the operation mode selected by the user based on the mode selection signal transmitted from the operation panel 220. Then, the switching unit 330 determines whether or not to execute the process by the second detection unit 13 with reference to, for example, a relationship table that defines the correspondence relationship between the operation mode and the character detection operation.

FIG. 18 is a diagram showing an example of a relationship table that defines the correspondence relationship between operation modes and character detection operations. In the relation table T1 shown in FIG. 18, when the “standard mode” is selected as the operation mode, the first detection unit 11 is activated, the second detection unit 13 is set to non-active, and the operation mode is the “character priority mode”. Is selected, it indicates that both the first detection unit 11 and the second detection unit 13 are activated.

When it is determined that the processing by the second detection unit 13 is to be executed, the switching unit 330 sends a control signal for activating the second detection unit 13 to the second processing unit 320, and the second detection unit If it is decided not to execute the processing by 13, the control signal for deactivating the second detection unit 13 is sent to the second processing unit 320. As a result, the active/non-active state of the second detection unit 13 is switched according to the operation mode selected by the user.

As described above, in the present embodiment, whether or not to execute the processing by the second detection unit 13 which requires a relatively long processing time is switched according to the operation mode selected by the user. Therefore, according to the present embodiment, which of the processing speed (or productivity) and the character separation performance (or the higher image quality of the character portion), which are difficult to achieve when generating the high compression PDF file FIm, is prioritized. Can be selected according to the user's wishes, and the user's dissatisfaction can be alleviated.

<First Modification>
In the second embodiment described above, when the “standard mode” is selected as the operation mode, the first detection unit 11 is active and the second detection unit 13 is non-active, but the operation mode is When the “standard mode” is selected, both the first detection unit 11 and the second detection unit 13 may be inactive. In this case, since the character is not detected from the multi-valued image Im1, the processing by the separating unit 14 is not executed, and the characters and the background included in the multi-valued image Im1 have the same compression method.

In order to carry out this modification, for example, in accordance with a predetermined operation of the user using the operation panel 220, for example, the first table corresponding to the “standard mode” in the relationship table T1 as shown in FIG. The state of the detection unit 11 may be changed from active to non-active. In addition, such a change of the relationship table T1 may be performed by a general user, or may be performed by only a specific user such as an administrator.

In this modified example, when the “standard mode” is selected, the image quality of the character may be degraded, but a portion other than the character in the pattern is not erroneously detected as the character. Therefore, for the user who attaches great importance to the image quality of the design, a preferable result can be obtained by selecting the “standard mode”.

<Second Modification>
Further, in the above-described second embodiment, the compression processing of the data stored in the HDD 213 and the decoding processing of the data read from the HDD 213 are performed regardless of the operation mode, but the “standard mode” is set as the operation mode. When selected, such a compression/decoding process may not be performed.

It is assumed that the user who selects the "standard mode" as the operation mode is a user who attaches importance to the processing speed (or productivity) and the image quality of the picture rather than the character separation performance (or the high image quality of the character portion). It As described above, the compression processing of the data stored in the HDD 213 and the decoding processing of the data read from the HDD 213 are useful for effectively using the capacity of the HDD 213, but the processing speed (or productivity) is emphasized. For the user, the increase in processing time due to such compression/decoding processing may cause dissatisfaction. Further, depending on the compression/decoding processing method, the image quality of the design may be deteriorated, which may cause dissatisfaction to the user who attaches importance to the image quality of the design.

In this modified example, when the “standard mode” is selected as the operation mode, the compression process of the data stored in the HDD 213 and the decryption process of the data read from the HDD 213 are not performed so that the “standard mode” is set. The above-mentioned dissatisfaction of the user who selects can be alleviated. Further, when the “standard mode” is selected as the operation mode, in addition to not performing the compression processing of the data stored in the HDD 213 and the decoding processing of the data read from the HDD 213, or not performing the compression/decoding processing. Instead of this, it is also effective to store and read data in the HDD 213 in page units.

<Third Modification>
Further, in the above-described second embodiment, whether to execute the process by the second detection unit 13 is switched according to the operation mode selected by the user. However, in addition to the operation mode, the processing speed is changed. May be selected by the user, and whether or not to execute the process by the second detection unit 13 may be switched depending on the combination of the operation mode and the processing speed selected by the user.

FIG. 19 is a diagram showing an example of an operation screen for accepting the selection of the processing speed in addition to the selection of the operation mode. The operation screen 190 shown in FIG. 19 has a configuration in which check boxes 191, 192 for selecting the processing speed for each operation mode are added to the operation screen 170 shown in FIG.

In the operation screen 190 shown in FIG. 19, the check box 191 corresponding to the "standard mode" is in a state where "medium speed" is selected by default. When the user touches the “standard mode” button 172 to select “standard mode” in the state of this default setting, as described above, the first detection unit 11 is active and the second detection unit 13 is non-active. Become. On the other hand, when the user selects “low speed” in the check box 191 corresponding to the “standard mode” and touches the “standard mode” button 172 to select “standard mode”, for example, the first detection unit 11 And the second detection unit 13 are both activated. When the user selects “high speed” in the check box 191 corresponding to the “standard mode” and touches the “standard mode” button 172 to select “standard mode”, for example, the first detection unit 11 and Both of the two detection units 13 become non-active.

Further, on the operation screen 190 shown in FIG. 19, the check box 192 corresponding to the “character priority mode” is in a state where “low speed” is selected by default. When the user touches the “character priority mode” button 173 and selects the “character priority mode” in the state of this default setting, as described above, both the first detection unit 11 and the second detection unit 13 are active. Becomes On the other hand, when the user selects “medium speed” in the check box 192 corresponding to the “character priority mode” and touches the “character priority mode” button 173 to select the “character priority mode”, for example, the first The detection unit 11 is active and the second detection unit 13 is non-active. When the user selects “high speed” in the check box 192 corresponding to the “character priority mode” and touches the “character priority mode” button 173 to select “character priority mode”, for example, the first detection unit Both 11 and the second detection unit 13 are non-active.

In the above, if the processing speed selected by the user is “high speed”, both the first detection unit 11 and the second detection unit 13 are made inactive, and the processing speed selected by the user is “medium speed”. If so, the first detection unit 11 is made active, the second detection unit 13 is made inactive, and if the processing speed selected by the user is “low speed”, the first detection unit 11 and the second detection unit 13 In this example, both are activated. However, the active/non-active state of the first detection unit 11 and the second detection unit 13 may be switched according to a relationship table that defines the correspondence relationship of the character detection operation with respect to the combination of the operation mode and the processing speed.

FIG. 20 is a diagram showing an example of a relationship table that defines the correspondence relationship of the character detection operation with respect to the combination of the operation mode and the processing speed. In the relationship table T2 shown in FIG. 20, when the "standard mode" is selected as the operation mode, the "low speed" is selected as the processing speed, and the "medium speed" is selected as the processing speed, It is shown that the first detector 11 is made active and the second detector 13 is made inactive. In this example, assuming that the user who attaches importance to the image quality of the pattern selects the “standard mode”, if the “standard mode” is selected as the operation mode and the “low speed” is selected as the processing speed, for example, The process of increasing the sharpness of the pattern by controlling the resolution of the. Therefore, by making the second detection unit 13 inactive, the time required for character detection is shortened, but the total processing time is lengthened.

Further, in the relationship table T2 shown in FIG. 20, when “character priority mode” is selected as the operation mode and “high speed” is selected as the processing speed, the same as when “medium speed” is selected as the processing speed. , The first detection unit 11 is made active, and the second detection unit 13 is made inactive. In this example, it is assumed that the user who selects the "character priority mode" wants to secure the image quality of the character to some extent even at a high speed, so that the "character priority mode" is selected as the operation mode and the "processing speed" is selected. When “high speed” is selected, the first detection unit 11 is activated to secure a certain character image quality, and the processing speed is increased by lowering the resolution of the pattern, for example, to reduce the total processing time. To shorten.

In the present modification, whether or not to execute the process by the second detection unit 13 is switched according to the combination of the operation mode selected by the user and the processing speed, so that the intention of the user is more faithful. The reflected processing becomes possible.

<Other modifications>
The first to fourth modified examples of the first embodiment described above are also applicable to the second embodiment.

[Third Embodiment]
Next, a third embodiment will be described. Similar to the second embodiment, the present embodiment is an example in which the function as the above-described image processing apparatus is realized in a multifunction machine, and the image read by the scanner engine 260 is the multivalued image Im1 to be processed. However, in the present embodiment, not only the purpose of generating the high compression PDF file FIm based on the image read by the scanner engine 260 (hereinafter, the case of performing image processing for this purpose is referred to as “during file generation”), but also the scanner It is also assumed that the plotter engine 270 (see FIG. 15) records a color image on a recording medium based on the image read by the engine 260 (hereinafter, the case where image processing is performed for this purpose is referred to as “during reproduction”). Then, in the present embodiment, the processing by the first detection unit 11 is switched between the time of file generation and the time of copy reproduction. Whether the file is generated or the copy is reproduced can be determined based on, for example, a job setting signal output when the user sets a job using the operation panel 220.

As described above, the first detection unit 11 uses the continuity and the pattern of black pixels and white pixels obtained by ternarizing the multivalued image Im1 to be processed to detect the edges that form a character. To detect. Therefore, the accuracy of edge detection by the first detection unit 11 can be controlled by the threshold value used when the multi-valued image Im1 is ternarized. Here, at the time of generating the file, as described above, in order to appropriately detect the edge of the low-contrast character such as the gray character included in the multi-valued image Im1, the threshold of the ternarization is used as the edge detection accuracy. It is desirable to set a threshold value at which On the other hand, at the time of copy reproduction, if the multi-valued image Im1 is ternarized using the same threshold value as that at the time of file generation, there is a problem that a portion that should be treated as a pattern from the viewpoint of image quality is also detected as a character edge. For example, when reproducing a copy, the area of black characters is reproduced in black (K) single color, but when the characters in a background with a relatively high density such as newspaper characters are reproduced in black single color, a gap with the background around the characters is generated. It becomes large and causes deterioration of image quality. Further, although there are few edge portions having extremely high contrast locally in the pattern, many edge portions having low contrast exist. For this reason, when the multi-valued image Im1 is ternarized using the same threshold value as that at the time of file generation during copy reproduction, it becomes a factor of image quality deterioration.

Therefore, in the present embodiment, as an example of switching the processing by the first detection unit 11, the threshold used when the multi-valued image Im1 is ternarized is switched between the file generation and the copy reproduction to thereby generate the file. At times, the edge detection accuracy of the first detection unit 11 is set higher than that at the time of copy generation, and at the time of copy reproduction, the edge detection accuracy of the first detection unit 11 is set lower than that at the time of file generation. That is, the threshold for ternarization is set so that the edge can be easily detected when the file is generated, and the threshold for ternarization is set so that the edge is less likely to be detected when the copy is reproduced as compared with the time when the file is generated. As a result, when the threshold for ternarization is fixed, it is possible to achieve both efficient compression at the time of file generation, which is in a trade-off relationship, and high image quality at the time of copy reproduction.

Further, in the present embodiment, the configuration of the first detection unit 11 is slightly different from the above-described first and second embodiments in consideration of the processing at the time of copy reproduction. That is, in addition to the edge detection by the same method as in the above-described first and second embodiments, the first detection unit 11 in the present embodiment determines whether the target pixel is a white background pixel, and It is determined whether the pixel is a pixel forming a halftone dot, and the determination result is comprehensively determined to detect the edge of the character from the multivalued image Im1 to be processed. In addition, in the present embodiment, in addition to the processing by the first detection unit 11, it is also determined (color determination) whether the target pixel is a chromatic block pixel or an achromatic block pixel. Then, at the time of copy reproduction, a process of converting the multivalued image Im1 to be processed into an image signal that can be processed by the plotter engine 270 is performed based on the determination result of the first detection unit 11 and the color determination result.

Hereinafter, a configuration example of the multifunction peripheral of this embodiment will be described. Note that, in the following, constituent elements that are common or correspond to those of the above-described first and second embodiments will be denoted by the same reference numerals, and redundant description will be omitted as appropriate. The hardware configuration of the multi-function peripheral of this embodiment can be the same as that of the multi-function peripheral 200 of the above-described second embodiment (see FIG. 15), and the description thereof will be omitted.

FIG. 21 is a block diagram showing a functional configuration example of the multifunction peripheral 200A of this embodiment. In the multifunction machine 200A of the present embodiment, as shown in FIG. 21, the first processing unit 310 includes a gamma correction unit 311 and a color determination unit 312 in addition to the first detection unit 11 and the edge enhancement unit 12. A data interface unit 313, a color processing/UCR unit 314, and a printer correction unit 315 are provided.

The gamma correction unit 311 performs a primary conversion process (gamma correction) on each color signal in order to adjust the gradation balance for each color of the multi-valued image Im1 to be processed (the image read by the scanner engine 260). In the present embodiment, the multivalued image Im1 to be processed is an RGB image signal represented by 8 bits for each RGB color, and the signal after conversion by the gamma correction unit 311 is a density linear (white is a signal value 0) RGB image signal. Shall be The multi-valued image Im1 that has been gamma-corrected by the gamma correction unit 311 is sent to the first detection unit 11, the edge enhancement unit 12, the color determination unit 312, and the data interface unit 313.

FIG. 22 is a block diagram showing a configuration example of the first detection unit 11 in this embodiment. As shown in FIG. 22, the first detection unit 11 in this embodiment includes an edge detection circuit 401, a white background detection circuit 402, a halftone dot detection circuit 403, and a comprehensive determination circuit 404.

The edge detection circuit 401 detects a character edge from the gamma-corrected multivalued image Im1 by the same method as the first detection unit 11 in the above-described first and second embodiments. That is, the edge detection circuit 401 ternarizes the multi-valued image Im1 (density linear) after gamma correction with two threshold values (th_w and th_b: th_w<th_b) to convert it into a black pixel, a white pixel, and a gray pixel. Divide. Then, by extracting a black line drawing or a white line drawing from the multi-valued image Im1 by pattern matching of the black connected pixels or the white connected pixels, the edge of the character included in the multi-valued image Im1 is detected.

At this time, the edge detection circuit 401 determines based on the job setting signal Y input from the operation panel 220 whether the file is generated or the copy is reproduced. Then, the edge detection circuit 401 switches the threshold values th_w and th_b used for ternarizing the multi-valued image Im1 after gamma correction between the file generation and the copy reproduction. For example, when the high density threshold th_b used during copy reproduction is th_b0 and the high density threshold th_b used during file generation is th_b1, the high density threshold th_b is switched so that th_b1<th_b0. As a result, it becomes easier to detect a low-contrast edge as a character edge during file generation, and it becomes difficult to detect a low-contrast edge as a character edge during copy reproduction. Further, when the low density threshold th_w used during copy reproduction is th_w0 and the low density threshold th_w used during file generation is th_w1, the low density threshold th_w is switched so that th_w1>th_w0. This makes it easier to detect an edge in the background having a relatively high density as a character edge during file generation, and makes it difficult to detect an edge in the background having a relatively high density as a character edge during copy reproduction. The threshold values th_w and th_b for ternarization may be switched by only one of the high-density threshold value th_b and the low-density threshold value th_w, or by both.

The edge detection circuit 401 detects the edge of the character included in the multi-valued image Im1 by the above processing, and outputs the result to the comprehensive determination circuit 404. The output of the edge detection circuit 401 is 1 bit per pixel, and the edge of the detected character is activated.

The white background detection circuit 402 determines whether the multivalued image Im1 after gamma correction is a white background or a non-white background, and outputs the result to the comprehensive determination circuit 404. The white background detection circuit 402, for example, binarizes the multi-valued image Im1 after gamma correction into a white pixel and a black pixel by binarizing the multi-valued image Im1 and separates the white pixel and the black pixel from each other. Then, the white background is determined. At this time, by controlling the sizes of the reference regions on the left, right, top, and bottom, it is possible to determine the edges of the characters having the desired line width or less as the white background and the edges of the characters exceeding the desired line width as the non-white background. it can. The output of the white background detection circuit 402 activates the white background.

The halftone dot detection circuit 403 performs halftone dot judgment (judgment based on a repeating pattern of peak/valley peak pixels in the image) for each pixel of the multivalued image Im1 after gamma correction, and the result is stored in the comprehensive judgment circuit 404. Output.

More specifically, the halftone dot detection circuit 403, for each pixel of the multi-valued image Im1 after gamma correction, has a matrix of predetermined M×M pixels, for example, as shown in FIGS. A matrix of 3×3 pixel size (M=3), a matrix of 4×4 pixel size (M=4) or a matrix of 5×5 pixel size (M=5) as shown in FIG. m ₀ (see FIG. 23 (a) ~ (c) ) for detecting whether or not a pole showing a peak or valley density changes from the concentration relationship between the surrounding pixels _m 1 ~m _i. In addition, the halftone dot detection circuit 403 uses the block B composed of N×N pixels (where N>M), for example, the block B composed of 9×9 pixel size (N=9) as shown in FIG. 24 as a unit. Divide the image.

Then, the halftone dot detection circuit 403 counts the number of pole pixels indicating peaks and the number of pole pixels indicating valleys for each block, and determines the number of pole pixels on the side with the larger count value as the number of pole pixels of the block. .. Thereafter, halftone detection circuit 403, the the relationship between the pole number of pixels P0 of the target block B ₀ shown in FIG. 25, each pole number of pixels P of the upper, lower, left and right diagonal of each surrounding blocks B ₁ .about.B ₈ surrounding the central pixel _{n 0} of the target block _{B 0} (see FIG. 24), or all of the pixels _n 0 _{~n 80} of the block of interest _B within ₀ determines whether belonging to the halftone dot region. The output of the halftone dot detection circuit 403 activates the halftone dot area.

The comprehensive determination circuit 404 comprehensively determines the determination result of the edge detection circuit 401, the determination result of the white background detection circuit 402, and the determination result of the halftone dot detection circuit 403 to determine whether or not the pixel of interest is a character. A determination result (character/non-character [pattern]) X1 indicating whether or not is output. That is, when the target pixel is active in the edge detection circuit 401, active in the white background detection circuit 402, and inactive in the halftone dot detection circuit 403, the comprehensive determination circuit 404 outputs a determination result X1 in which the target pixel is a character. .. When a file is generated, the determination result X1 output by the comprehensive determination circuit 404 corresponds to the character edge detection result in the above-described first and second embodiments. The determination result X1 output by the comprehensive determination circuit 404 is sent to the edge enhancement unit 12 and the data interface unit 313. Note that when generating a file, the output of the edge detection result 401 may be used as the character edge detection result instead of the determination result X1 output by the comprehensive determination circuit 404.

The edge enhancement unit 12 uses the determination result X1 of the comprehensive determination circuit 404 of the first detection unit 11 to perform the process of enhancing the edges of the characters included in the multi-valued image Im1 to generate the edge enhanced image Im2. The edge emphasized image Im2 generated by the edge emphasizing unit 12 is sent to the data interface unit 313.

The color determination unit 312 determines whether the pixel block of interest (4×4 pixels) is a chromatic color block or an achromatic color block in the multi-valued image Im1 after gamma correction. FIG. 26 is a block diagram showing a configuration example of the color determination unit 312. The color determination unit 312 includes a maximum value calculation unit 411, a first comparison unit 412, a chromatic pixel counter 413, and a second comparison unit 414, as shown in FIG. 26, for example.

The maximum value calculation unit 411 calculates the maximum value d[i,j] (=ΔRGB) of the absolute value of the RGB difference for each pixel of the multi-valued image Im1 after gamma correction.

The first comparison unit 412 compares the maximum value d[i,j] calculated by the maximum value calculation unit 411 with a predetermined pixel determination threshold th_pix. Then, a pixel whose maximum value d[i,j] exceeds the pixel determination threshold th_pix is determined as a chromatic pixel, and a pixel whose maximum value d[i,j] is less than or equal to the pixel determination threshold th_pix is determined as an achromatic pixel.

The chromatic pixel counter 413 counts, for each pixel block of a predetermined number of pixels (for example, 4 lines×4 pixels) included in the multi-valued image Im1 after gamma correction, the number C1 of chromatic pixels included in the pixel block. ..

The second comparison unit 414 compares the number C1 of chromatic pixels in the pixel block counted by the chromatic pixel counter 413 with a predetermined block determination threshold th_blc. A pixel block in which the number C1 of chromatic pixels exceeds the block determination threshold th_blc is determined to be a chromatic block, and a pixel block in which the number C1 of chromatic pixels is equal to or less than the block determination threshold th_blc is determined to be an achromatic block. /Achromatic) X2 is output. The determination result X2 output by the second comparison unit 414 is sent to the data interface unit 313.

The data interface unit 313 generates a multi-valued image Im1 after gamma correction, a determination result X1 of the first detection unit 11 (corresponding to a character edge detection result), and an edge-enhanced image Im2 generated by the edge enhancement unit 12 when a file is generated. , And a determination result X2 of the color determination unit 312 when temporarily stored in the HDD 213. At the time of file generation, as in the second embodiment described above, the second detection unit 13 and the separation unit 14 of the second processing unit 320 read necessary data from the HDD 213 as needed and perform the above-described processing. Then, the file generation unit 15 compresses the character image Im3 and the background image Im4 separated from the multi-valued image Im1 by the separation unit 14 by different methods and then integrates them to obtain a high compression PDF file corresponding to the multi-valued image Im1. Generate FIm.

On the other hand, at the time of copy reproduction, the multi-valued image Im1 after gamma correction, the determination result X1 of the first detection unit 11 and the determination result X2 of the color determination unit 312 are transmitted via the data interface unit 313 to the color processing/UCR unit 314. Sent to.

The color processing/UCR unit 314 selects color processing or UCR processing based on the determination results X1 and X2 for each pixel or pixel block, and the plotter engine 270 processes the gamma-corrected multivalued image Im1 (RGB image signal). Convert to an image signal that can be.

FIG. 27 is a block diagram showing a configuration example of the color processing/UCR unit 314. The color processing/UCR unit 314 includes a first color processing unit 421, a UCR unit 422, a second color processing unit 423, and a selector 424, as shown in FIG. 27, for example. The multi-valued image Im1 (RGB image signal) after gamma correction output for each pixel from the data interface unit 313 is input in parallel to the first color processing unit 421 and the second color processing unit 423.

The first color processing unit 421 performs color reproduction processing for non-black characters. The first color processing unit 421 performs RGB→CMY conversion by, for example, a 3×3 matrix calculation represented by the following expression (1) in order to realize color reproduction faithful to the multivalued image Im1 to be processed. To do. At this time, in order to improve the accuracy of color reproduction, for example, the RGB space may be divided and the matrix calculation may be performed for each area.

The UCR unit 422 performs black signal generation and black signal replacement from CMY obtained by the color reproduction processing of the first color processing unit 421 according to the following equations (2) to (5). The c, m, y, Bk image signals obtained by the processing of the UCR unit 422 are input to the selector 424. In addition, α in the following formula (2) is an adjustment parameter that takes a value of 0 to 1.0.
Bk=α×min (C, M, Y) (2)
c=C-Bk (3)
m=M-Bk (4)
y=Y-Bk (5)

On the other hand, the second color processing unit 423 performs color reproduction processing for black characters. The color reproduction process of the second color processing unit 423 is a process of calculating a signal (Bk0 image signal) corresponding to luminance from an RGB image signal, for example. The Bk0 image signal (C=M=Y=0) obtained by the processing of the second color processing unit 423 is input to the selector 424.

The selector 424 selects the c, m, y, Bk image signals for non-black characters or the Bk0 image signal for black characters based on the determination results X1, X2 output from the data interface unit 313, and the printer correction unit. Output to 315. That is, the selector 424 selects the Bk0 image signal and outputs it to the printer correction unit 315 if the pixel of interest is a character according to the determination result X1 and is achromatic according to the determination result X2. On the other hand, if the pixel of interest is a fire letter based on the determination result X1, or if it is a chromatic color based on the determination result X2, the c, m, y, Bk image signals are selected and output to the printer correction unit 315.

The printer correction unit 315 subjects the image signal output from the color processing/UCR unit 314 to gamma correction processing and dither processing according to the characteristic peculiar to the plotter engine 270, and outputs the image signal to the plotter engine 270. The plotter engine 270, which is a transfer printing unit, attaches a recording material such as toner or ink to the recording medium based on the image signal output from the printer correction unit 315. As a result, a color image corresponding to the multivalued image Im1 to be processed (the image read by the scanner engine 260) is recorded on the recording medium.

As described above, in the present embodiment, the processing by the first detection unit 11 is switched between the time of file generation and the time of copy reproduction. More specifically, the threshold value of the ternarization performed by the edge detection circuit 401 is switched between the file generation and the copy reproduction. Therefore, according to the present embodiment, when the threshold for ternarization is fixed, it is possible to achieve both efficient compression at the time of file generation, which is in a trade-off relationship, and high image quality at the time of copy reproduction. it can.

<First Modification>
In the third embodiment described above, when the file is generated, the separation unit 14 of the second processing unit 320 detects the determination result X1 (corresponding to the character edge detection result) of the first detection unit 11 and the second determination result X1. It is assumed that the multi-valued image Im1 is separated into the character image Im3 and the background image Im4 using the detection result of the character area by the detection unit 13. However, the determination result X1 of the first detection unit 11 may not be used as it is for the processing by the separation unit 14, but the determination result X1 may be corrected and then used.

FIG. 28 is a block diagram showing a functional configuration example of a multifunction peripheral 200A' of the present modification. As shown in FIG. 28, in the multifunction peripheral 200A' of the present modification, a character area correction unit 321 is added to the second processing unit 320. Other than that, the configuration is the same as that of the third embodiment (see FIG. 21).

The reasons for wanting to correct the determination result X1 of the first detection unit 11 include, for example, the following points. That is, the determination result X1 of the first detection unit 11 is the result of detecting the edge of the character, but considering efficient compression at the time of file generation, not only the edge of the character but also the area inside the character becomes a character. It is desirable to handle. On the other hand, the background of a character is a background near the character and its surroundings, and a texture difference becomes conspicuous depending on the compression method used for each. Therefore, it is desirable to correct the determination result X1 of the first detection unit 11 so that the background of the character is determined again as a non-character (pattern). The area inside the character is configured to be detected as a character by the global analysis in the second detection unit 13. However, by correcting the determination result X1 of the first detection unit 11, the second detection is performed. There is also an effect of compensating for the detection omission of the part 13.

FIG. 29 is a block diagram showing a configuration example of the character area correction unit 321. For example, as shown in FIG. 29, the character area correction unit 321 includes a luminance signal calculation circuit 431, a binarization circuit 432, a line buffer 433, and a mask correction circuit 434.

The luminance signal calculation circuit 431 calculates the signal L corresponding to the luminance signal according to the following equation (6) for each pixel of the multi-valued image Im1 (RGB image signal) after gamma correction. In addition, a, b, and c in the following equation (6) are coefficients that have been previously obtained by experiments.
L=a×R+b×G+c×B (6)

The binarization circuit 432 performs binarization (L≧th_L→black pixel, L<th_L→white pixel) processing on the signal L calculated by the luminance signal calculation circuit 431 with a predetermined threshold th_L. The line buffer 433 holds the processing result of the binarization circuit 432 and inputs the information necessary for the processing in the mask correction circuit 434 to the mask correction circuit 434.

The mask correction circuit 434 determines that the pixel of interest is a black pixel, and the pixel of which the determination result X1 of the first detection unit 11 is active is in the 7×7 mask centered on the pixel of interest. If there is even one, the determination result X1 for the pixel of interest is made active, and if not, the determination is made non-active again. FIG. 30 is a diagram showing an example of a mask used by the mask correction circuit 434. The pixel at the center of the mask shown in FIG. 30 (pixel with hatching in the figure) is the target pixel.

By the processing of the character area correction unit 321, the determination result X1 of the first detection unit 11 is corrected so that the inside of the character is active and the background of the character is inactive. Then, when the separating unit 14 performs the process of separating the multi-valued image Im1 into the character image Im3 and the background image Im4, the separating unit 14 uses the determination result X1 corrected by the character area correcting unit 321. The separation performance can be further improved.

<Second Modification>
Further, in the above-described third embodiment, as an example of switching the processing of the first detection unit 11 between file generation and copy reproduction, the threshold value for ternarization in the edge detection circuit 401 of the first detection unit 11 is set. The example of switching has been described. However, in addition to the switching of the threshold for ternarization, the processing of the first detection unit 11 may be switched by another method. For example, the signal of the multi-valued image Im1 input to the first detection unit 11 may be switched between when the file is generated and when the copy is reproduced.

Specifically, for example, at the time of copy reproduction, the G signal of the RGB image signal which is the multi-valued image Im1 after gamma correction is input to the first detection unit 11 in order to improve the efficiency of the processing by the first detection unit 11. .. On the other hand, at the time of generating the file, the signal L corresponding to the luminance signal is calculated from the RGB image signal which is the multi-valued image Im1 after gamma correction according to the above equation (6), and the signal L is input to the first detection unit 11. To do. When the processing by the first detection unit 11 is performed based on the G signal, the processing can be made efficient, but the response to the green portion is low, so that it is difficult to detect the edge of the green character. On the other hand, if the processing by the first detection unit 11 is performed based on the signal L corresponding to the luminance signal, it becomes easier to detect the edge of the green character.

When it is desired to positively detect not only the edge of the green character but also the edge of the yellow character, the maximum value max(R, G, B) of RGB is used instead of the signal L corresponding to the luminance signal at the time of file generation. It may be configured to input to the first detection unit 11.

<Third Modification>
Further, in the above-described third embodiment, the white background detection circuit 402 and the halftone dot detection circuit 403 of the first detection unit 11 are used only for detecting the character edge. However, the output of the white background detection circuit 402 and the output of the halftone dot detection circuit 403 may be used in the processing in the separation unit 14, that is, when the multi-valued image Im1 is separated into the character image Im3 and the background image Im4. .. Further, the compression method may be switched when the file generation unit 15 compresses the image based on the output of the white background detection circuit 402 and the output of the halftone dot detection circuit 403. In order to realize such a configuration, the output of the white background detection circuit 402 and the output of the halftone dot detection circuit 403 are sent to the data interface unit 313 independently of the determination result X1 of the first detection unit 11, and the HDD 213 is operated. You can save it temporarily in.

<Other modifications>
The above-described first to third embodiments and their modifications can be appropriately combined and implemented. For example, in the configuration of the first embodiment or the second embodiment, the processing by the first detection unit 11 (for example, a threshold for ternarization) is switched depending on whether the file is generated or the copy is reproduced as in the third embodiment. You may

Although specific embodiments and modifications of the present invention have been described above, the above-described embodiments show one application example of the present invention. The present invention is not limited to the above-described embodiments as they are, and can be embodied by making various modifications and changes within a range not departing from the gist of the invention in an implementation stage.

DESCRIPTION OF SYMBOLS 1 Image processing apparatus 11 1st detection part 12 Edge emphasis part 13 2nd detection part 14 Separation part 15 File generation part 200 Multifunction machine 213 HDD
330 Switching unit Im1 multi-valued image Im2 edge enhanced image Im3 character image Im4 background image FIm highly compressed PDF file

Japanese Patent No. 3088010 Japanese Patent No. 4471202

Claims (15)

A first detection unit for detecting a character edge from a multi-valued image;
An edge enhancement unit that enhances an edge of a character included in the multi-valued image using the detection result of the first detection unit;
A second detection unit for detecting a character area from the multi-valued image in which the edges of the character are emphasized;
An image processing apparatus comprising: a separation unit that separates the multi-valued image into a character image and a background image using at least the detection result of the second detection unit. The separation unit separates the multi-valued image into a character image and a background image using the detection result of the first detection unit and the detection result of the second detection unit. Image processing device. The separating unit detects at least one of a color ground character region and a white character region in the multi-valued image using the detection result of the second detecting unit and the multi-valued image, and detects the detected region. Separating the multi-valued image into a character image and a background image by using the detection result of the first detection unit excluding the edge of the character included in the above and the detection result of the second detection unit. Item 2. The image processing device according to item 2. The second detection unit acquires a connected component from a binarized image obtained by binarizing the multi-valued image in which the edges of the characters are emphasized, and the size of the circumscribed rectangle is predetermined among the acquired connected components. The image processing device according to claim 2, wherein the character region is detected by deleting a connected component that is less than a size and integrating the remaining connected components. A first detection unit for detecting a character edge from a multi-valued image;
A second detection unit for detecting a character area from the multi-valued image;
Using the detection result of the second detection unit and the multi-valued image, at least one of the area of the color ground character and the area of the white character in the multi-valued image is detected, and the character included in the detected area is detected. An image including a separation unit that separates the multi-valued image into a character image and a background image by using the detection result of the first detection unit excluding edges and the detection result of the second detection unit. Processing equipment. The separation unit integrates the detection result of the first detection unit and the detection result of the second detection unit to separate the character image,
The image processing apparatus according to claim 1, wherein a difference between the multi-valued image and the character image is separated as the background image. 7. The file generation unit according to claim 1, further comprising a file generation unit that compresses the character image and the background image by different methods and then integrates them to generate a compressed image file corresponding to the multi-valued image. The image processing device described. A switching unit for switching whether or not to execute the process by the second detection unit according to the operation mode selected by the user;
The separation unit separates the multi-valued image into the character image and the background image using at least the detection result of the second detection unit when the process by the second detection unit is executed. The multi-valued image is separated into the character image and the background image by using the detection result of the first detection unit when the process by the second detection unit is not executed. The image processing device according to claim 1. The switching unit switches whether or not to execute the process by the second detection unit according to a combination of an operation mode selected by the user and a processing speed selected by the user. Image processing device. Further comprising a storage unit for storing at least the multi-valued image,
The second detection unit detects a character area from the multi-valued image read from the storage unit,
10. The separation unit separates the multi-valued image read from the storage unit into a character image and a background image using at least the detection result of the second detection unit. The image processing device according to 1. The first detection unit generates a compressed image file corresponding to the multi-valued image and records an image corresponding to the multi-valued image on a recording medium. The image processing apparatus according to any one of claims 1 to 10, which switches processing for detecting a. The processing by the first detection unit includes processing for ternarizing the multi-valued image,
The first detection unit ternarizes the multi-valued image depending on whether a compressed image file corresponding to the multi-valued image is generated or an image corresponding to the multi-valued image is recorded on a recording medium. The image processing apparatus according to claim 11, wherein a threshold value for switching is switched. The first detection unit uses a threshold value that is easier to detect an edge when generating a compressed image file corresponding to the multi-valued image than when an image corresponding to the multi-valued image is recorded on a recording medium. 13. The image processing apparatus according to claim 12, wherein the multi-valued image is ternarized. An image processing method executed in an image processing apparatus, comprising:
A first detection step of detecting an edge of a character from a multi-valued image,
An edge emphasis step of emphasizing an edge of a character included in the multi-valued image using the detection result of the first detection step;
A second detection step of detecting a character area from the multi-valued image in which edges of characters are emphasized,
An image processing method including a separation step of separating the multi-valued image into a character image and a background image using at least the detection result of the second detection step. On the computer,
The function of the first detection unit that detects the edge of a character from a multi-valued image,
A function of an edge enhancement unit that enhances edges of characters included in the multi-valued image using the detection result of the first detection unit;
A function of a second detection unit that detects a character area from the multi-valued image in which edges of characters are emphasized,
A program for realizing a function of a separation unit that separates the multi-valued image into a character image and a background image by using at least the detection result of the second detection unit.

Images