JP2023030420A - Image processing apparatus, image processing method, and program - Google Patents

Abstract

To improve processing efficiency in image processing till obtaining processed image data in accordance with a result of detecting a specific object image.SOLUTION: An image processing apparatus 70 which performs image processing on a region including a specific object image included in high-resolution image data includes: an image resizing unit 70a which reduces the high-resolution image data at a predetermined magnification to obtain low-resolution resized image data; a specific object image detection unit 70c which detects a first position information group related to the specific object image from the low-resolution resized image data obtained by the image resizing unit 70a; and an image processing unit 70g which converts, when the specific object image detection unit 70c detects the first position information group, the first position information group into a second position information group in accordance with an inverse number value of the predetermined magnification, and processes only an image of a region included in the second position information group in the high-resolution image data, to generate processed image data.SELECTED DRAWING: Figure 2

Description

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

2. Description of the Related Art Conventionally, an image forming apparatus has an image processing function. For example, image data used for copy output is obtained by detecting a specific object image from image data in advance and obtaining processed image data after performing processing on the area of the object image in the image data. .
Conventionally, in image processing for obtaining processed image data,
(1) generating low-resolution image data by thinning pixels from high-resolution image data;
(2) Next, detect the position of the object image from the low-resolution image data,
(3) Further, using the position of the object image detected in (2), high-resolution image data is read out to detect the detailed position of the object image (analyze shape, numbers, etc.),
(4) Next, process the high-resolution image data according to the position of the detected object image.
Image processing was performed in this order.
Since the processes (1) to (4) are performed in this order, there is a problem that the processing efficiency in the image processing is lowered in order to obtain the processed image data to be used for the copy output.

The invention described in Patent Document 1 aims to improve the recognition accuracy of images for primary judgment without increasing the bus capacity and performance of an image sensor that processes high-resolution camera images in real time according to the camera. was intended to As image data for primary determination (for object detection), reduced image data v1 obtained by thinning an image captured by a high-resolution camera (original image data v) at a constant ratio is used. When the detection processing unit detects the movement or change of the car based on v1, the detection processing unit sends a transfer request for the original image data v for secondary judgment of the frame buffer unit to the transfer control unit, and also sets the detection pixel area to The indicated detection position address is sent to the detailed analysis unit. Based on this transfer request, the detailed analysis section analyzes the shape, number, etc. of the detection position address portion of the original image data v received. It also discloses a case where (v−v1) is held in the frame buffer unit, and a case where the address generation unit transfers only the original image data of the analysis target range to the detailed analysis unit based on the detection position address.
That is, Japanese Patent Application Laid-Open No. 2002-200002 discloses a configuration for performing processing on a low-resolution image and processing on a high-resolution image for the purpose of performing object image detection processing. However, the problem that the processing efficiency in the image processing until obtaining the image data to be used for copy output is lowered has not been solved.

An embodiment of the present invention has been made in view of the above, and its object is to improve processing efficiency in image processing up to obtaining processed image data according to the detection result of a specific object image.

In order to solve the above-mentioned problems, the invention according to claim 1 is an image processing apparatus for performing image processing on a region including a specific object image included in high-resolution image data, wherein the high-resolution image data includes: an image size scaling unit that reduces image data by a predetermined scaling ratio to obtain low-resolution scaled image data; and the specific object from the low-resolution scaled image data obtained by the image size scaling unit. a specific object image detection unit for detecting a first position information group related to an image; an image processing unit that converts a first group of positional information into a second group of positional information and processes only an image of an area included in the second group of positional information in the high-resolution image data to generate processed image data; , is provided.

According to the present invention, it is possible to prevent a decrease in processing efficiency in image processing until processed image data is obtained according to the detection result of a specific object image.

1 is a block diagram showing the hardware configuration of an image forming apparatus including an image processing apparatus according to an embodiment of the invention; FIG. 1 is a functional block diagram of an image processing device according to an embodiment of the present invention; FIG. 4 is a flow chart showing an outline of image processing by the image processing apparatus according to the embodiment of the present invention; 3A shows a front-stage image processing unit used for reading image data, and FIG. 3B shows a rear-stage image processing unit used for writing image data; FIG. 4 is a flow chart showing operation of image processing used for object image detection by the image processing apparatus according to the embodiment of the present invention. 5 is a flow chart showing the operation of image processing according to detection results by the image processing apparatus according to the embodiment of the present invention; 4 is a flow chart showing the operation of image processing according to the face image detection result by the image processing apparatus according to the embodiment of the present invention. 5 is a flow chart showing the operation of image processing according to detection results by the image processing apparatus according to the embodiment of the present invention; 4 is a flow chart showing details of image processing for detecting an object image in a specific m×n size image by the image processing apparatus according to the embodiment of the present invention. FIG. 4 is an explanatory diagram showing a specific object image and an image used for copy output by the image processing apparatus according to the embodiment of the present invention;

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will now be described in detail with reference to embodiments shown in the drawings.
The present invention has the following configuration in order to improve processing efficiency in image processing up to obtaining processed image data according to the detection result of a specific object image.
That is, the image processing apparatus of the present invention is an image processing apparatus that processes an image of a region containing a specific object image included in high-resolution image data, and processes high-resolution image data at a predetermined scaling factor. An image size scaling unit for reducing and obtaining low-resolution scaled image data, and detecting a first position information group related to a specific object image from the low-resolution scaled image data obtained by the image size scaling unit. and when the specific object image detection unit detects the first position information group, the first position information group is converted into the second position information group according to the reciprocal value of the predetermined scaling factor. and an image processing unit that processes only the image of the area included in the second position information group in the high-resolution image data to generate processed image data.
By providing the above configuration, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the specific object image.
The features of the invention described above will be explained in detail with reference to the following drawings. However, unless there is a specific description, the components, types, combinations, shapes, relative arrangements, and the like described in this embodiment are merely illustrative examples and not intended to limit the scope of the present invention. .
The features of the present invention described above will be described in detail below with reference to the drawings.

<Hardware Configuration of Image Forming Apparatus>
FIG. 1 is a block diagram showing the hardware configuration of an image forming apparatus including an image processing apparatus according to an embodiment of the invention.
FIG. 1 is a hardware configuration diagram of an MFP, which is an example of an image forming apparatus. As shown in FIG. 1, an MFP (Multifunction Peripheral/Product/Printer) includes a controller 20, a short-range communication circuit 60, an engine control section 30, an operating device 40, and a network I/F50.
Among them, the controller 20 includes a CPU 25a, a system memory (MEM-P) 22, a north bridge (NB) 25b, a south bridge (SB) 24, an ASIC (Application Specific Integrated Circuit) 26, a storage unit, which are the main parts of the computer. , a local memory (MEM-C) 27, an HDD controller 28, and an HD 29, which is a storage unit.
Among these, the CPU 25 a is a control unit that performs overall control of the MFP 10 . The NB 25b is a bridge for connecting the CPU 25a, the MEM-P 22, the SB 24, and the AGP bus 21. The NB 25b is a memory controller that controls reading and writing with respect to the MEM-P 22, a PCI (Peripheral Component Interconnect) master, and an AGP target. have

The MEM-P 22 consists of a ROM 22a, which is a memory for storing programs and data for realizing each function of the controller 20, and a RAM 22b, which is used as a drawing memory for developing programs and data and for memory printing. The program stored in the RAM 22b is configured to be provided by being recorded in a computer-readable recording medium such as a CD-ROM, CD-R, DVD, etc. as a file in an installable format or an executable format. You may

The SB 24 is a bridge for connecting the NB 25b with PCI devices and peripheral devices. The ASIC 26 is an image processing IC (Integrated Circuit) having hardware elements for image processing, and serves as a bridge connecting the AGP bus 21, PCI bus 23, HDD controller 28 and MEM-C 27, respectively. This ASIC 26 includes a PCI target and AGP master, an arbiter (ARB) that forms the core of the ASIC 26, a memory controller that controls the MEM-C 27, and multiple DMACs (Direct Memory Access Controllers) that rotate image data using hardware logic. , and a PCI unit for transferring data between the scanner section 31 and the printer section 32 via the PCI bus 23 . The ASIC 26 may be connected to a USB (Universal Serial Bus) interface or an IEEE 1394 (Institute of Electrical and Electronics Engineers 1394) interface.

MEM-C27 is a local memory used as a copy image buffer and code buffer. The HD 29 is a storage for accumulating image data, accumulating font data used for printing, and accumulating forms. The HD 29 controls reading or writing of data to or from the HD 29 under the control of the CPU 25a. The AGP bus 21 is a bus interface for graphics accelerator cards proposed for speeding up graphics processing, and by directly accessing the MEM-P 22 with high throughput, the graphics accelerator card can be speeded up. .
The short-range communication circuit 60 also includes a short-range communication circuit 60a. The short-range communication circuit 60 is a communication circuit for NFC, Bluetooth, or the like.

Further, the engine control section 30 is composed of a scanner section 31 and a printer section 32 . The operation device 40 also displays a current setting value, a selection screen, etc., and displays a panel display unit 40a such as a touch panel for receiving an input from an operator, and a setting value of conditions related to image formation such as a setting condition of density. An operation panel 40b is provided which includes a numeric keypad for accepting a copy start instruction, a start key for accepting a copy start instruction, and the like. The controller 20 controls the entire MFP, such as drawing, communication, and input from the operation device 40 . The scanner section 31 or the printer section 32 includes an image processing section such as error diffusion and gamma conversion.

Note that the MFP can switch and select the document box function, the copy function, the printer function, and the facsimile function in sequence using the application switching key of the operation device 40 . The document box mode is set when the document box function is selected, the copy mode is set when the copy function is selected, the printer mode is set when the printer function is selected, and the facsimile mode is set when the facsimile mode is selected.

A network I/F 50 is an interface for data communication using a communication network. The short-range communication circuit 60 and network I/F 50 are electrically connected to the ASIC 26 via the PCI bus 23 .
The controller 20 accepts a command issued through the operation device 40 or a command issued by an external device, and executes job processing instructed by the command.
Among these, the portion related to the printer portion 32 from the engine control portion 30 is the portion related to the present invention, and the description will focus on this point. The plotter control section 30p, which is a part of the engine control section 30, will be described below.

<Functional block diagram>
FIG. 2 is a functional block diagram of the image processing apparatus according to the embodiment of the invention.
The image processing device 70 shown in FIG. 2 processes an image of a specific object image area included in high-resolution image data.
The image processing apparatus 70 includes units that function by activating the plotter control unit 30p described above, that is, an image size scaling unit 70a, a specific object image detection unit 70c, a determination unit 70f, and an image processing unit 70g.

The image size scaling unit 70a reduces the high resolution image data by a predetermined scaling factor to obtain low resolution scaled image data.
The specific object image detection unit 70c includes a position detection unit 70d, and detects a first position information group related to a specific object image from the low-resolution scaled image data obtained by the image size scaling unit 70a.
When the specific object image detection unit 70c detects the first position information group, the image processing unit 70g converts the first position information group into a second position information group according to the reciprocal value of a predetermined scaling factor, Only the image of the area included in the second position information group in the high-resolution image data is processed to generate processed image data.

When the specific object image is a face image, the specific object image detection unit 70c detects the first position information group related to the face image, and the image processing unit 70g detects the first position information group related to the face image. If so, the image of the area included in the second position information group related to the face image detected by the position detection unit is processed.

The specific object image detection unit 70c includes a QR code (registered trademark) detection unit 70e, extracts data represented by the QR code (registered trademark) when the specific object image is a QR code (registered trademark), and Detecting the first position information group related to the QR code (registered trademark), the image processing unit 70g detects the copy prohibited data in the data represented by the QR code (registered trademark). The image of the area included in the second position information group related to is processed.

The image size scaling unit 70a reduces the size of image data to a fixed size of m×n.

The image size scaling unit 70a has a neural network used for inference processing when detecting a specific object image from scaled image data, and the fixed size is a size that can be processed by the neural network.

Each function of the embodiments described above may be implemented by one or more processing circuits. Here, the "processing circuit" in this specification means a processor programmed by software to perform each function, such as a processor implemented by an electronic circuit, or a processor designed to perform each function described above. ASICs (Application Specific Integrated Circuits), DSPs (digital signal processors), FPGAs (field programmable gate arrays) and devices such as conventional circuit modules.

<Outline of image processing>
FIG. 3 is a flow chart showing an outline of image processing by the image processing apparatus according to the embodiment of the present invention. The overall processing flow will be described with reference to FIG.
In step S10, the plotter control section 30p causes the scanner section 31 to scan the document, and develops the obtained high-resolution image data in the RAM 22b. The plotter control unit 30p performs the following two parallel processes on the image data developed in the RAM 22b.
In step S15, the plotter control unit 30p performs image processing for the image data used for copy output, and outputs the image data used for copy output according to the output resolution of the printer unit 32. FIG.

In step S20, the specific object image detection unit 70c performs image processing for detecting an object image. In object image detection processing, it is often possible to perform detection processing even with high-resolution image data, but the processing time increases in proportion to the number of pixels.
Although it depends on the detection processing algorithm, the detection accuracy is more effective when the resolution is increased. However, when the resolution exceeds a certain level, the effect of improving the detection accuracy does not increase significantly. On the other hand, the processing time increases as the resolution increases.
Therefore, depending on the purpose, the image processing used for object image detection detects the presence or absence of a specific object image in the image data obtained by scanning, or specifies the position of the area of the object image.

In step S25, the image processing unit 70g performs image processing on the image data used for copy output according to the detection result in the post-stage image processing unit 90b shown in FIG. Processing is performed on the material.
As an example, if it is desired to prohibit copying of a document on which specific QR code (registered trademark) information is written, the image data used for copy output is replaced with a specific image pattern, mosaic processing, or black painting processing. See FIGS. 10A and 10B for explanatory diagrams of such processing.
In addition, if it is desired not to copy the face photograph for the purpose of personal information protection, for example, the face area of the driver's license image is subjected to mosaic processing or blackening processing. See FIGS. 10(c) and 10(d) for explanatory diagrams of such processing.
In step S30, the plotter control section 30p sends the processed image data to the printer section 32 and prints the image data on a recording medium.

<Configuration example of image processing unit>
FIG. 4 is a diagram showing a configuration example of an image processing unit included in the image processing apparatus according to the embodiment of the present invention, where (a) is a front-stage image processing unit used for reading image data, and (b) is an image data is a diagram showing a post-stage image processing unit used for writing of .
The pre-stage image processing unit 90a used for scanning shown in FIG. 4A may be provided in the scanner, and performs only image processing during reading.
In the post-stage image processing unit 90b used for writing shown in FIG. 4B, not only printing on a recording medium but also displaying on a display device such as a monitor may be performed. image processing only.

The MFP shown in FIG. 1 includes both a front-stage image processing unit 90a used for reading and a rear-stage image processing unit 90b used for writing.
Scanned images of MFPs and scanners and printed images of printers are configured so that they can be temporarily stored in a memory or an external storage device after image processing is completed for reuse of the image data.
As shown in FIGS. 4A and 4B, the image processing section of the MFP includes a front-stage image processing section 90a that mainly processes image data of the reading device, and a rear-stage image processing section 90a that mainly performs processing for the writing device. The configuration includes an image processing unit 90b.

<Image processing for reading device>
As shown in FIG. 4A, the front-stage image processing unit 90a includes a line-to-line correction processing unit 91a, an image area separation processing unit 91b, a scaling processing unit 91c, a γ conversion processing unit 91d, a filter processing unit 91e, a color conversion It has a configuration including a processing unit 91f.
The line-to-line correction processing unit 91a is a process for correcting a line shift between RGB caused by a difference in mounting position of each RGB of the color CCD. For example, when the B (blue) line is used as a reference, a process of correcting the amount of line deviation between R (red) and B and between G (green) and B is performed.
The image area separation processing unit 91b generates image area separation information X such as a pattern portion, a character portion, and a halftone dot portion from the features of the image data. This separation information is used in the filter processing and γ conversion processing of the post-stage image processing unit 90b to set image processing parameters suitable for each region.

The variable magnification processing unit 91c performs processing for converting the read resolution to a desired resolution.
The γ conversion processing unit 91d performs conversion processing mainly for density adjustment, and generally uses a method called lookup table conversion.
The filter processing unit 91e performs filter processing calculations for the purpose of MTF correction, sharpening, and smoothing.
The color conversion processing unit 91f performs conversion processing to a common color space for storage in the memory controller 80. FIG. Also, when automatic color determination is selected, K (monochrome) data is generated from RGB data. This monochrome data is simply binarized and packed into 8-bit data as 8-pixel data. RGBK 4ch image data is transferred to the memory controller 80 . In normal color data reading, RGB 3-channel image data is transferred. In addition, if there is an application that requires separation information, such as a PDF file, 4ch data of RGB data and image area separation information is transferred to the memory controller 80 .

<Image processing for writing device>
As shown in FIG. 4B, the post-stage image processing section 90b includes a color conversion processing section 92b and a gradation conversion processing section 92c.
The color conversion processing unit 92b performs processing for converting the accumulated data RGB as input data into the color space of the output device, for example, the CMYK color space. In the case of color output, the above processing is performed for each of the RGB components, and in the case of monochrome output, monochrome data is generated from RGB.
Gradation conversion by binarization will be described using a fixed threshold when the writing device is 1-bit and can output up to 2 gradations. When a binary image is desired, the gradation conversion processing unit 92c shown in FIG. It is sent to the image processing section 90b.
Here, to simplify the explanation, an example of fixed threshold processing will be given. value.
0 ≤ pixel data < 0 if 128 is true
1 if 128 ≤ pixel data ≤ 255 is true

Next, gradation conversion by fixed threshold quaternary conversion when the writing device can output up to 2 bits and 4 levels will be described. When a 4-level image is desired, the gradation conversion processing unit 92c converts an image of 8 bits for each of CMYK and 256 gradations into 4-level image data of 4 gradations, and sends the data as image data b to the post-stage image processing unit 90b.
Here, to simplify the explanation, as an example of fixed threshold value processing, the pixel data of the input image of the processing section is quaternarized under the following conditions.
0 ≤ pixel data < 0 if 64 is true
1 if 64 ≤ pixel data < 128 is true
2 if 128 ≤ pixel data < 192 is true
3 if 192 ≤ pixel data ≤ 255 is true
In addition to the simple binarization described above, gradation conversion processing such as error diffusion processing and dither processing, which are suitable for photographic originals and have excellent gradation properties, may be performed, although they are not described here.

<Image processing used for object image detection>
FIG. 5 is a flow chart showing the operation of image processing used for object image detection by the image processing apparatus according to the embodiment of the present invention.
In step S50, the image size scaling unit 70a performs scaling processing.
For example, if the reading resolution of the scanner is 600 dpi and the output resolution of the printer is 600 dpi, image processing for copying is performed at 600 dpi. If so, it is necessary to speed up the detection processing speed.
Therefore, in step S50, the image size scaling unit 70a converts the 600 dpi scanned image, which is a high-resolution image, into a 200 dpi image, which is a low-resolution image, according to the scaling factor (1/3) in the scaling process. to shrink.

Next, in step S55, the specific object image detection unit 70c performs specific object image detection processing on this image.
Next, when switching the processing of the post-stage image processing unit 90b simply based on whether or not the specific object image is included in the scanned image, in step S60, the specific object image detection unit 70c detects whether or not the specific object image exists. is notified to the post-stage image processing unit 90b.
When switching the processing of the post-stage image processing unit 90b for a specific object image region, position information of the object is passed as a detection result. Depending on the purpose, the image of the detection area is also transferred to the post-stage image processing unit 90b (for example, in the case of QR code (registered trademark) detection).
At this time, the presence/absence information of a specific object image may be notified as position information.
For example, if the position information of four points in a rectangular area containing a specific object image is all (0, 0), then the object image is in an undetected state. If it is set as position information, it may be assigned such that it is determined that a specific object image has been detected.

<Image processing according to detection results>
FIG. 6 is a flow chart showing the operation of image processing according to detection results by the image processing apparatus according to the embodiment of the present invention.
In step S55 shown in FIG. 5, the specific object image detection unit 70c detects whether or not an image of a specific stamp is included as the specific object image. A case in which information indicating that there is an image is notified will be described.

First, in FIG. 6, in step S105, the determination unit 70f determines whether or not a specific object image has been detected.
When a specific object image is detected, in step S110, the image processing unit 70g replaces the specific object image area in the image data used for copy output with a specific pattern image, and the document information cannot be read. processed as
On the other hand, if the specific object image is not detected, in step S115, the plotter control section 30p passes the image data used for the copy output of the preceding image processing section 90a to the succeeding image processing section 90b as it is.

<Image processing according to face image detection results>
FIG. 7 is a flow chart showing the operation of image processing according to the face image detection result by the image processing apparatus according to the embodiment of the present invention.
A case in which a face image is detected in a document and mosaic processing is performed on the area of the face image will be described with reference to FIG.
Incidentally, as a face outline detection process, a "face recognition gate system" described in Japanese Patent No. 6180906 has been reported.
In facial contour detection processing, facial feature points are detected as a list, and these points represent the feature shape of the face. Recognize as facial contours.

In step S105 in FIG. 6, the determination unit 70f performs face image detection processing as the specific object image detection processing.
In FIG. 7, in step S150, the specific object image detection unit 70c determines whether or not the image data obtained by scanning the document image by the scanner unit 31 includes a face image area. Here, when a face image is detected (S150, Yes), the process proceeds to step S155, and position information (x1, y1) (x2, y2) (x3, y3) (x4) of four points of a rectangular region surrounding the face region is detected. , y4) as a position information group.

Since the position information group of these four points corresponds to the low-resolution image, in step S160, the image processing unit 70g converts them into a copy output image (high-resolution image) according to the reciprocal value of the scaling factor. is converted into a four-point position information group corresponding to .
For example, if a low-resolution image used to detect a specific object image has a resolution of 300 dpi and an output image for copying (high-resolution image) has a resolution of 600 dpi, then four points of position information group (x1, y1 ) (x2, y2) (x3, y3) (x4, y4) are respectively doubled (scaling ratio) to represent the position of the specific object image corresponding to the copy output image (high resolution image) A four-point position information group (X1, Y1) (X2, Y2) (X3, Y3) (X4, Y4) is obtained.

In step S165, the image processing unit 70g processes the face image area (the area surrounded by (X1, Y1) (X2, Y2) (X3, Y3) (X4, Y4) in the copy output image (high-resolution image). ) with the pattern image. That is, the calculated rectangular area is subjected to mosaic processing or filling processing, and the processed image is passed to the post-stage image processing section 90b as image data to be used for copy output. For example, see FIG. 10(b).
On the other hand, if the plotter control unit 30p determines in step S150 that the scan image does not include the facial image area (S150, No), the process proceeds to step S170 and outputs the image data as it is.

<Image processing according to detection results>
FIG. 8 is a flow chart showing the operation of image processing according to detection results by the image processing apparatus according to the embodiment of the present invention.
Image processing when a QR code (registered trademark) is detected as a specific object image will be described with reference to FIG.
In step S105 of FIG. 6, the QR code (registered trademark) detection unit 70e performs detection processing of a specific object image.
Regarding the QR code (registered trademark), a "two-dimensional code" described in Japanese Patent No. 2938338 has been reported.
The smallest unit (black and white square) that constitutes the QR Code (registered trademark) is called a cell, and the QR Code (registered trademark) is represented by a combination of cells, and includes a position detection pattern (finder pattern), timing pattern, error correction It consists of format information having information such as level and mask number, data, and error correction code (Reed-Solomon code).

Further, as for the reading of the QR code (registered trademark), a "two-dimensional code reader" described in Japanese Patent No. 28679042 has been reported.
In reading a two-dimensional code, for example, the two-dimensional code is extracted from an image captured by a CCD and decoded. The image picked up at this time is converted into a binary image for extracting the code from the image (gray image) picked up as it is. Code information read from a QR code (registered trademark) may include code information and a code indicating copy prohibition.

In FIG. 8, in step S205, the QR code (registered trademark) detection unit 70e determines whether the scanned image data includes a QR code (registered trademark). If the image data contains a QR code (registered trademark) (S205, Yes), the process proceeds to step S210. , the process proceeds to step S225.
If the image data contains a QR code (registered trademark) (S205, Yes), the process proceeds to step S210, and the QR code (registered trademark) detector 70e extracts an image containing a QR code (registered trademark) from the image data. Since the image data obtained by cutting out the area is also transferred, the code information of the QR code (registered trademark) of the image of this area is read.

In step S215, the determination unit 70f determines whether or not the code information of the QR code (registered trademark) is a code indicating copy prohibition.
If the code information of the QR code (registered trademark) is a code indicating copy prohibition (S215, Yes), the process proceeds to step S220, and the image processing unit 70g specifies image data to be used for copy output by the preceding image processing unit 90a. is replaced with the pattern image of , and passed to the post-stage image processing unit 90b. Refer to FIG. 10B for the print result in this case.
On the other hand, if the code information of the QR code (registered trademark) is not a code indicating copy prohibition (S215, No), the process proceeds to step S225.
If the image data does not contain a QR code (registered trademark) or if the information of the QR code (registered trademark) is not copy-prohibited, the image data used for the copy output of the preceding image processing unit 90a is copied and output as it is. image data to be used in the subsequent image processing unit 90b. Refer to FIG. 10A for the print result in this case.

<Image processing for detecting a specific object image using an m×n size image>
FIG. 9 is a flowchart showing details of image processing for detecting a specific object image using a specific m×n size image by the image processing apparatus according to the embodiment of the present invention.
A case where a deep learning neural network (DNN) is used for object image detection processing will be described with reference to FIG.
The process of detecting a particular object image using many neural networks assumes a particular image size. Assume here that the image size is, for example, 256×256. The size of the scanned image data varies depending on the actual document size, but it is necessary to reduce the image size to an image size that can be input to the neural network (DNN).

First, in step S255, the image size scaling unit 70a calculates a scaling factor.
For example, when copying an A4 portrait document at 600 dpi output, the size of the scanned image is 4960×7016, so it is necessary to reduce this to an image size of 256×256 that can be input to the DNN. In this case, the scaling factor in the main scanning direction is 256/4960=5.14%, and the scaling factor in the subscanning direction is 256/7016=3.64%.
In step S260, the image size scaling unit 70a sets the scaling factor to the scaling processing unit 91c to reduce the size of the image data.
The size of the image data used in the object image detection processing of the post-stage image processing unit 90b varies depending on the image size of the adopted DNN. Different sizes, for example 64×64 and 192×192. When the MFP has functions for a plurality of purposes, the neural network (DNN) is switched according to the purpose of the object image detection processing, and at the same time, the size of the detected image is also switched.
For example, the size is switched to 64×64 for QR code (registered trademark) detection processing and 192×192 for face detection processing.
In step S265, the specific object image detection unit 70c performs detection processing of a specific object image. That is, the first position information group related to the specific object image is detected from the low-resolution magnified image data.
In step S270, the image processing unit 70g performs conversion processing of position information represented by coordinates according to the reciprocal value of the scaling factor calculated in step S255. That is, the first positional information group is converted into the second positional information group according to the reciprocal value of the scaling factor.

<Specific object image and image used for copy output>
10A to 10D are explanatory diagrams showing specific object images and images used for copy output by the image processing apparatus according to the embodiment of the present invention.
FIG. 10(a) is an unprocessed image of a driver's license image 101 including a facial photograph 101a as a specific object image. FIG. 10(b) is an image used for copy output after processing of a driver's license image 103 including a processed image 103a obtained by processing the image of the facial photograph 101a area of FIG. 10(a).
FIG. 10(c) is an image before processing of the back side image 105 of the membership card of the Traffic Safety Association including the QR code (registered trademark) 105a as the specific object image. FIG. 10(d) is a copy output after processing of the back side image 107 of the membership card of the Traffic Safety Association including the processed image 107a obtained by processing the image of the area of the QR code (registered trademark) 105a in FIG. 10(c). It is an image used for

<Effects of this embodiment>
According to this embodiment, high-resolution image data is reduced by a predetermined scaling ratio to obtain low-resolution scaled image data, and a specific object image is obtained from the obtained low-resolution scaled image data. When the related first position information group is detected and the first position information group is detected, the first position information group is converted into the second position information group according to the reciprocal value of the predetermined scaling factor, and the high resolution Only the image of the area included in the second position information group in the image data is processed to generate processed image data.
As a result, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the specific object image.
According to this embodiment, the image processing for high-resolution image data in step S15 and the specific object image detection processing for low-resolution image data in step S20 can be performed in parallel. It is possible to improve processing efficiency when copying image data.

<Summary of Actions and Effects of Mode Examples of the Present Embodiment>
<First aspect>
The image processing apparatus 70 of this aspect is an image processing apparatus 70 that performs image processing on a region including a specific object image included in high-resolution image data, and converts high-resolution image data to a predetermined scaling factor. and first positional information relating to a specific object image from the low-resolution scaled image data obtained by the image size scaling unit 70a. When the specific object image detection unit 70c detects the first position information group, the first position information group is transferred to the second position according to the reciprocal value of the predetermined scaling factor. and an image processing unit 70g that converts the data into information groups and processes only the image of the area included in the second position information group in the high-resolution image data to generate processed image data.
According to this aspect, high-resolution image data is reduced by a predetermined scaling ratio to obtain low-resolution scaled image data, and a specific object image is obtained from the obtained low-resolution scaled image data. Detecting a first group of position information, and when detecting the first group of position information, converting the first group of position information into a second group of position information according to the reciprocal value of a predetermined scaling factor, and obtaining a high-resolution image Processed image data is generated by processing only the image of the area included in the second position information group in the data.
As a result, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the specific object image.

<Second aspect>
In the image processing device 70 of this aspect, when the specific object image is a face image, the specific object image detection unit 70c detects the first position information group related to the face image,
The image processing unit 70g is characterized in that when the first position information group related to the face image is detected, the image of the area included in the second position information group related to the face image detected by the position detection unit is processed. do.
According to this aspect, when the specific object image is a face image, the first position information group related to the face image is detected, and when the first position information group related to the face image is detected, the first position information group related to the face image is detected. The image of the area included in the second position information group is processed.
As a result, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the face image.

<Third aspect>
When the specific object image is a QR code (registered trademark), the specific object image detection unit 70c of this aspect extracts data represented by the QR code (registered trademark), 1 position information group is detected, and if the data represented by the QR code (registered trademark) contains copy prohibited data, the image processing unit 70g detects the data included in the second position information group related to the QR code (registered trademark). It is characterized by processing the image of the area where the
According to this aspect, when the specific object image is a QR code (registered trademark), the data represented by the QR code (registered trademark) is extracted, and the first position information group related to the QR code (registered trademark) is obtained. If the data represented by the QR code (registered trademark) contains copy prohibited data, the image of the area included in the second position information group related to the QR code (registered trademark) is processed.
As a result, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the QR code (registered trademark).

<Fourth Aspect>
The image size scaling unit 70a of this aspect is characterized by reducing the size of image data to a fixed size of m×n.
According to this aspect, the size of the image data is reduced to a fixed size of m×n.
As a result, it is possible to improve the processing efficiency when detecting the first position information group related to a specific object image from low-resolution magnified image data.

<Fifth aspect>
The image size scaling unit 70a of this aspect includes a neural network used for inference processing when detecting a specific object image from scaled image data, and the fixed size is a size that can be processed by the neural network. and
According to this aspect, the neural network used for inference processing when detecting a specific object image from scaled image data is provided, and since the fixed size is a size that can be processed by the neural network, It is possible to improve processing efficiency when detecting the first position information group related to a specific object image from data.

<Sixth Aspect>
The image processing method of this aspect is an image processing method by an image processing device that performs image processing on a region including a specific object image included in high-resolution image data, and the high-resolution image data is processed by a predetermined image processing method. an image size scaling step of reducing by a scaling factor to obtain low-resolution scaled image data; and first position information relating to a specific object image from the low-resolution scaled image data obtained by the image size scaling step. a specific object image detecting step for detecting groups, and when the specific object image detecting step detects the first positional information group, the first positional information group is converted to the second positional information group according to the reciprocal value of a predetermined scaling factor. and processing only the image in the area included in the second position information group in the high-resolution image data to generate processed image data.
According to this aspect, high-resolution image data is reduced by a predetermined scaling ratio to obtain low-resolution scaled image data, and a specific object image is obtained from the obtained low-resolution scaled image data. Detecting a first group of position information, and when detecting the first group of position information, converting the first group of position information into a second group of position information according to the reciprocal value of a predetermined scaling factor, and obtaining a high-resolution image Processed image data is generated by processing only the image of the area included in the second position information group in the data.
As a result, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the specific object image.

<Seventh Aspect>
A program of this aspect is a program for causing a processor to function as an image processing method for performing image processing on a region including a specific object image included in high-resolution image data, an image size scaling step for obtaining low-resolution scaled image data by reducing the image data by a predetermined scaling ratio; and a specific object image from the low-resolution scaled image data obtained by the image size scaling step. and when the specific object image detection step detects the first position information group, the first position information group according to the reciprocal value of a predetermined scaling factor into the second position information group, and processing only the image of the area included in the second position information group in the high-resolution image data to generate processed image data. It's a program.
According to this aspect, high-resolution image data is reduced by a predetermined scaling ratio to obtain low-resolution scaled image data, and a specific object image is obtained from the obtained low-resolution scaled image data. Detecting a first group of position information, and when detecting the first group of position information, converting the first group of position information into a second group of position information according to the reciprocal value of a predetermined scaling factor, and obtaining a high-resolution image Processed image data is generated by processing only the image of the area included in the second position information group in the data.
As a result, it is possible to improve the processing efficiency in the image processing until the processed image data is obtained according to the detection result of the specific object image.

Reference Signs List 10 Image forming apparatus 20 Controller 22a ROM 22b RAM 25a CPU 28 HDD controller 29 HD 30 Engine control unit 30p Plotter control unit 31 Scanner unit 32 Printer unit 40 Operation device 40a Panel display unit 40b Operation panel 60 Short-range communication circuit 60a Short-range communication circuit 70 Image processing device 70a Image size scaling unit 70c Specification Object image detection unit 70d Position detection unit 70e QR code (registered trademark) detection unit 70f Judgment unit 70g Image processing unit 80 Memory controller 90a Pre-stage image processing unit 90b Post-stage image processing Section 91a Inter-line correction processing unit 91b Image area separation processing unit 91c Variable magnification processing unit 91d Conversion processing unit 91e Filter processing unit 91f Color conversion processing unit 92b Color conversion processing unit , 92c... Gradation conversion processing unit 101... Driver's license image 101a... Face photograph 103... Driver's license image 103a... Processed image 105... Membership card reverse side image 105a... QR code (registered trademark) 107 ...Membership card reverse side image, 107a...Processed image

Japanese Patent Application Laid-Open No. 2006-127195

Claims (7)

An image processing device that performs image processing on a region containing a specific object image included in high-resolution image data,
an image size scaling unit that reduces the high resolution image data by a predetermined scaling factor to obtain low resolution scaled image data;
a specific object image detection unit that detects a first position information group related to the specific object image from the low-resolution scaled image data obtained by the image size scaling unit;
When the specific object image detection unit detects the first positional information group, the first positional information group is converted into a second positional information group according to the reciprocal value of the predetermined scaling factor, and the high-resolution and an image processing unit that processes only an image of an area included in the second position information group in image data to generate processed image data. the specific object image detection unit, when the specific object image is a face image, detects a first position information group related to the face image;
The image processing unit according to claim 1, wherein when the first position information group related to the face image is detected, the image processing unit processes the image of the area included in the second position information group related to the face image. image processing device. The specific object image detection unit, when the specific object image is a QR code (registered trademark), extracts data represented by the QR code (registered trademark), 1 position information group is detected,
The image processing unit processes an image of an area included in a second position information group related to the QR code (registered trademark) when copy prohibited data is included in the data represented by the QR code (registered trademark). The image processing apparatus according to Claim 2, characterized by: 2. The image processing apparatus according to claim 1, wherein the image size scaling unit reduces the size of the image data to a fixed size of m*n. The image size scaling unit includes a neural network used for inference processing when detecting the specific object image from the scaled image data,
5. The image processing apparatus according to claim 4, wherein said fixed size is a size that can be processed in a neural network. An image processing method by an image processing device for performing image processing on a region including a specific object image included in high-resolution image data,
an image size scaling step of reducing the high-resolution image data by a predetermined scaling factor to obtain low-resolution scaled image data;
a specific object image detection step of detecting a first position information group related to the specific object image from the low-resolution scaled image data obtained by the image size scaling step;
When the specific object image detection step detects the first position information group, the first position information group is converted into a second position information group according to the reciprocal value of the predetermined scaling factor, and the high-resolution and an image processing step of processing only an image of an area included in the second position information group in image data to generate processed image data. A program for causing a processor to function as an image processing method for performing image processing on a region containing a specific object image included in high-resolution image data,
the processor,
an image size scaling step of reducing the high-resolution image data by a predetermined scaling factor to obtain low-resolution scaled image data;
a specific object image detection step of detecting a first position information group related to the specific object image from the low-resolution scaled image data obtained by the image size scaling step;
When the specific object image detection step detects the first position information group, the first position information group is converted into a second position information group according to the reciprocal value of the predetermined scaling factor, and the high-resolution and an image processing step of processing only the image of the area included in the second position information group in the image data to generate processed image data.

Images