JP2001188901A - Image processor, image processing method, and computer- readable recording medium with recorded program making computer implement same method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image after excellent density conversion free of image deterioration such as data absence even when the number of gradations which are converted is small and to perform fast density conversion which makes good use of the priority of an SIMD type processor. SOLUTION: A feature quantity extraction part 602 extracts a feature quantity from image data, vector data of output pixels are determined according to the feature quantity and pixel data of the respective pixels forming the image data and stored in a register 605, and a selector 607 selects vector data obtained through vector-added dither processing 606 or the vector data stored in the register 605 according to the feature quantity.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a SIMD (Single Instructed) which converts input image data read from an image reading unit to the output density of an image output unit.
Tion Multiple Data stream
The present invention relates to an image processing apparatus for performing density conversion using an m) type processor, an image processing method, and a computer-readable recording medium storing a program for causing a computer to execute the method.

[0002]

2. Description of the Related Art Conventionally, when an input image density is converted according to an output image density, a scaling process is generally used. However, if the number of output gradations is reduced, even if data is interpolated by scaling calculation, the number of output bits is small, so the interpolated data is almost rounded, and as a result, data is discarded. Increased, causing moiré and image deterioration.

For this reason, there is known a conventional technique for performing density conversion without such data drop. For example, Japanese Patent Application Laid-Open No. 7-37081 (Prior Art 1) discloses an image density conversion method in which the connectivity between a processing pixel and a neighboring pixel is represented by a numerical value, and a gradation conversion method is changed according to the magnitude of the numerical value. An apparatus is disclosed.

Japanese Patent Laying-Open No. 11-32210 (Prior Art 2) discloses a configuration in which data is replaced in a matrix and then a scaling process is performed, thereby performing an interpolation operation without dropping data at an edge portion. The disclosed image processing apparatus is disclosed. Further, in addition to these prior arts, there is a known prior art which performs density conversion while performing data interpolation so that an edge portion becomes smooth by pattern matching with a neighboring pixel and a neighboring pixel of a target pixel to be subjected to density conversion. Have been.

[0005]

However, the calculation of connectivity and the replacement of data in a matrix performed in these prior arts are not suitable for the processing of the SIMD type processor, so that the processing cannot be speeded up. There is a problem.

For example, when obtaining the maximum value of each pixel data in a 5 × 5 matrix input by a scanner,
ASIC (Application Specific)
On the Integrated Circuit, processing for sequentially comparing data of five lines in the main scanning direction is performed simultaneously, and finally, the maximum value obtained for each line is compared in the sub-scanning direction, and the processing is completed in six clocks. However, in the case of the SIMD type processor, there is a problem that a processing step of 25 instructions is required to perform a brute force comparison, and a processing delay occurs.

Further, in the SIMD type processor, for example, when comparing several patterns in which I / Os are arranged, a brute force comparison is performed. Doing so results in a cumulative number of processing steps.

According to the present invention, in order to solve the above-mentioned problem of the prior art, even if the number of gradations to be converted is a small value,
An image processing apparatus, an image processing method, and an image processing method capable of acquiring a good image after density conversion without image deterioration such as data omission, and performing high-speed density conversion by taking advantage of the advantage of the SIMD type processor; It is another object of the present invention to provide a computer-readable recording medium on which a program for causing a computer to execute the method is recorded.

[0009]

Means for Solving the Problems To solve the above-mentioned problems,
In order to achieve the object, an image processing apparatus according to the first aspect of the present invention performs density conversion of input image data read from an image reading unit using a SIMD type processor so as to match output density of an image output unit. In the image processing apparatus, determining means for determining characteristics of an image from input image data, and a first output pixel based on a determination result by the determining means and a pixel value of each pixel forming the image data.
Determining means for determining the vector data, a dither table storing a predetermined threshold value and second vector information, and the second vector data stored in the dither table or the first vector determined by the determining means. Selecting means for selecting any of the vector data based on the judgment result by the judging means; and gradation for performing a gradation process based on the vector data selected by the selecting means and a threshold value stored in the dither table. And processing means.

According to the first aspect of the present invention, a feature of an image is determined from input image data, and a first output pixel is determined based on the determination result and a pixel value of each pixel forming the image data. Is determined based on the determination result, either of the second vector data or the first vector data stored in the dither table, and based on the selected vector data and the threshold value stored in the dither table. Tone processing,
Even if the number of gradations to be converted is a small value, it is possible to obtain a good image after density conversion without image deterioration such as data omission, and high-speed density conversion utilizing the superiority of SIMD type processor Can be performed.

According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the selecting unit is configured to determine the second vector data stored in the dither table by the determining unit. The first vector data determined is preferentially selected.

According to the second aspect of the present invention, the first vector data determined by the determining means is preferentially selected over the second vector data stored in the dither table. Adjustment of the dot arrangement and conversion of the dot arrangement according to the characteristics of the image can be easily performed, and a high-quality density-converted image can be obtained.

According to a third aspect of the present invention, in the image processing apparatus according to the first or second aspect,
The gradation processing means can select vector data having a vector size of 0 when the quantization value is 3 or more.

According to the third aspect of the present invention, when the quantization value is 3 or more, the vector data having the vector size of 0 can be selected. The following dot arrangement conversion or dot conversion with good solid filling is easily performed in a small number of steps, so that a good density conversion image can be obtained.

According to a fourth aspect of the present invention, there is provided an image processing method for performing density conversion of input image data read from an image reading unit using a SIMD type processor so as to be adapted to an output density of an image output unit. In the processing method, a determining step of determining characteristics of an image from input image data; and a first output pixel based on a determination result of the determining step and a pixel value of each pixel forming the image data.
And a selecting step of selecting either the second vector data stored in the dither table or the first vector data determined in the determining step based on the determination result in the determining step. And a gradation processing step of performing gradation processing based on the vector data selected in the selection step and the threshold value stored in the dither table.

According to the present invention, the feature of the image is determined from the input image data, and the first output pixel is determined based on the determination result and the pixel value of each pixel forming the image data. Is determined based on the determination result, either of the second vector data or the first vector data stored in the dither table, and based on the selected vector data and the threshold value stored in the dither table. Tone processing,
Even if the number of gradations to be converted is a small value, it is possible to obtain a good image after density conversion without image deterioration such as data omission, and high-speed density conversion utilizing the superiority of SIMD type processor Can be performed.

According to a fifth aspect of the present invention, in the image processing method according to the fourth aspect, the selecting step is performed by the determining step more than the second vector data stored in the dither table. The first vector data determined is preferentially selected.

According to the fifth aspect of the present invention, the first vector data determined by the determining means is preferentially selected over the second vector data stored in the dither table. Adjustment of the dot arrangement and conversion of the dot arrangement according to the characteristics of the image can be easily performed, and a high-quality density-converted image can be obtained.

The image processing method according to the sixth aspect of the present invention is the image processing method according to the fourth or fifth aspect,
In the gradation processing step, vector data having a vector size of 0 when the quantization value is 3 or more can be selected.

According to the sixth aspect of the present invention, when the quantization value is 3 or more, the vector data having the vector size of 0 can be selected. The following dot arrangement conversion or dot conversion with good solid filling is easily performed in a small number of steps, so that a good density conversion image can be obtained.

According to a seventh aspect of the present invention, there is provided a recording medium on which a program for causing a computer to execute the method according to any one of the fourth to sixth aspects is recorded, so that the program is machine-readable. This makes it possible to realize the operations of claims 4 to 6 by a computer.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of an image processing apparatus, an image processing method, and a computer-readable recording medium storing a program for causing a computer to execute the method will be described below with reference to the accompanying drawings. Will be described in detail. In the present embodiment, a case will be described in which the present invention is applied to a complex digital multifunction peripheral such as a copier, a printer, a scanner, and a facsimile.

First, the principle of the image processing apparatus according to the present embodiment will be described. FIG. 1 is a block diagram functionally showing the configuration of the image processing apparatus according to the embodiment of the present invention. In FIG. 1, the image processing apparatus is formed from the following five units.

More specifically, as shown in FIG. 1, this image processing apparatus controls an image data control unit 100, an image reading unit 101 for reading image data, and an image memory for storing images to control image data. An image memory control unit 102 for writing / reading image data, an image processing unit 103 for performing image processing such as processing / editing on image data, and an image writing unit 104 for writing image data to transfer paper or the like.

Each unit includes an image data control unit 1
00, an image reading unit 101, an image memory control unit 102, an image processing unit 103
And the image writing unit 104 are connected to the image data control unit 100, respectively. Note that the density conversion of the binary image according to the present invention is performed by the image processing unit 103.

(Image Data Control Unit 100) The image data control unit 100 includes (1) data compression processing (primary compression) for improving data bus transfer efficiency, and (2) conversion of primary compressed data into image data. Transfer processing, (3) image synthesis processing (image data from a plurality of units can be synthesized, and also includes synthesis on a data bus), and (4) image shift processing (main scanning and sub-scanning) Image shift), (5) image area expansion processing (image area can be enlarged by an arbitrary amount to the periphery), (6) image scaling processing (for example, fixed scaling of 50% or 200%) ,
(7) parallel bus interface processing, (8) serial bus interface processing (interface with a process controller 211 described later),
(9) Format conversion processing of parallel data and serial data, (10) interface processing with the image reading unit 101, (11) interface processing with the image processing unit 103, and the like.

(Image Reading Unit 101) The image reading unit 101 comprises (1) reading processing of reflected light of a document by an optical system, and (2) CCD (Charge Coupled).
Device: charge-coupled device), conversion into an electric signal, (3) digitization by an A / D converter,
(4) Shading correction processing (processing for correcting illuminance distribution unevenness of a light source), (5) scanner γ correction processing (processing for correcting density characteristics of a reading system), and the like.

(Image Memory Control Unit 102) The image memory control unit 102 includes (1) interface control processing with a system controller, (2) parallel bus control processing (interface control processing with a parallel bus), and (3) network Control processing, (4) serial bus control processing (control processing of a plurality of external serial ports), (5) internal bus interface control processing (command control processing with the operation unit), (6) local bus control processing (system controller) RO to activate
M, RAM, font data access control processing),
(7) memory module operation control processing (memory module write / read control processing, etc.);
(8) Access control processing for memory modules (processing for arbitrating memory access requests from multiple units), (9) Data compression / expansion processing (to reduce the amount of data for effective use of memory) Processing),
(10) Image editing processing (data clear in memory area,
Image data rotation processing, image synthesis processing on a memory, etc.).

(Image Processing Unit 103) The image processing unit 103 includes (1) shading correction processing (processing for correcting unevenness in illuminance distribution of a light source), and (2) scanner γ correction processing (processing for correcting density characteristics of a reading system). ),
(3) MTF correction processing, (4) smoothing processing, (5) arbitrary scaling processing in the main scanning direction, (6) density conversion (γ conversion processing: corresponding to density notch), (7) simple multi-value processing, (8) Simple binarization processing, (9) error diffusion processing, (10) dither processing, (11) dot arrangement phase control processing (rightward dot,
Leftward dot), (12) isolated point removal processing, (13) image area separation processing (color determination, attribute determination, adaptive processing), (14)
Perform density conversion processing.

The image processing unit 103 supplies image data to a plurality of element processors, respectively.
SIMD-type processing is performed to cause each element processor to execute the same operation in parallel.

(Image Writing Unit 104) The image writing unit 104 includes (1) edge smoothing processing (jaggy correction processing), (2) correction processing for dot rearrangement, and (3) pulse control processing of image signals. And (4) format conversion between parallel data and serial data.

(Hardware Configuration of Digital MFP)
Next, a hardware configuration when the image processing apparatus according to the present embodiment forms a digital multifunction peripheral will be described. FIG. 2 is a block diagram illustrating an example of a hardware configuration of the image processing apparatus according to the present embodiment.

In the block diagram of FIG. 2, the image processing apparatus according to the present embodiment includes a reading unit 201, a sensor board unit 202, and an image data control unit 2.
03, an image processor 204, a video data control unit 205, and an imaging unit (engine) 206. The image processing apparatus according to the present embodiment includes a process controller 211, a RAM 212, and a ROM 213 via a serial bus 210.

Here, the image processor 204
Is a processor that performs image processing calculations using a SIMD type processor.
Data is exchanged with the outside via the.

The image processing apparatus according to this embodiment includes an image memory access control unit 221 and a facsimile control unit 224 via a parallel bus 220, and further includes an image memory access control unit 221.
Module 222, a system controller 231, a RAM 232, and a ROM 2
33 and an operation panel 234.

Here, the relationship between each of the above components and each of the units 100 to 104 shown in FIG. 1 will be described. That is, the reading unit 201 and the sensor board
The function of the image reading unit 101 shown in FIG. 1 is realized by the unit 202. Similarly, the function of the image data control unit 100 is realized by the image data control unit 203. Similarly, the image processor 204
Realizes the function of the image processing unit 103.

Similarly, the video / data control unit 205
The image writing unit 104 is realized by the image forming unit (engine) 206. Similarly, the image memory access control unit 221 and the memory module 2
The image memory control unit 102 is realized by 22.

Next, the contents of each component will be described. The reading unit 201 that optically reads a document includes a lamp, a mirror, and a lens, and condenses the reflected light of the lamp irradiation on the document to a light receiving element by the mirror and the lens.

A light receiving element, for example, a CCD is a sensor
The image data that is mounted on the board unit 202 and converted into an electric signal by the CCD is converted into a digital signal and then output (transmitted) from the sensor board unit 202.

The image data output (transmitted) from the sensor board unit 202 is transmitted to an image data control unit 203.
Is input (received). Functional device (processing unit)
The transmission of image data between the data buses is controlled entirely by the image data control unit 203.

The image data control unit 203 transfers data between the sensor board unit 202, the parallel bus 220, and the image processor 204, and processes the image data with the process controller 21 for the image data.
1 and a system controller 231 that controls the entire image processing apparatus. RAM2
Reference numeral 12 is used as a work area of the process controller 211, and the ROM 213 stores a boot program and the like of the process controller 211.

The image data output (transmitted) from the sensor board unit 202 is transmitted to an image data control unit 203.
(Transmit) to the image processor 204 via
Then, the signal deterioration due to the quantization into the optical system and the digital signal (referred to as the signal deterioration of the scanner system) is corrected and output (transmitted) to the image data control unit 203 again.

The image memory access control unit 221 comprises:
It controls writing / reading of image data to / from the memory module 222. In addition, the parallel bus 220
Controls the operation of each component connected to. Also, RAM
232 is used as a work area of the system controller 231, and the ROM 233 stores a boot program and the like of the system controller 231.

The operation panel 234 inputs a process to be performed by the image processing apparatus. For example, the type of process (copying, facsimile transmission, image reading, printing, etc.), the number of processes, and the like are input. Thereby, the input of the image data control information can be performed.

Next, the read image data is stored in the memory module 222 and reused.
There are jobs that are not stored in the memory module 222, and each case will be described. As an example of storing in the memory module 222, when copying a plurality of sheets of one document, the reading unit 201 is operated only once, and the image data read by the reading unit 201 is stored in the memory module 222. There is a method of reading stored image data a plurality of times.

As an example in which the memory module 222 is not used, when only one document is copied, the read image data may be reproduced as it is. Therefore, the memory module 222 by the image memory access control unit 221 is used. You do not need to access to.

First, when the memory module 222 is not used, the data transferred from the image processor 204 to the image data controller 203 is returned from the image data controller 203 to the image processor 204 again. The image processor 204 performs image quality processing for converting luminance data by the CCD in the sensor board unit 202 into area gradation.

The image data after the image quality processing is transferred from the image processor 204 to the video data controller 205. The post-processing relating to the dot arrangement and the pulse control for reproducing the dots are performed on the signal changed to the area gradation, and then the reproduced image is formed on the transfer paper in the image forming unit 206.

Next, a description will be given of the flow of image data in the case where additional processing such as rotation in the image direction, image synthesis, and the like are performed when the image is stored in the memory module 222 and the image is read out. The image data transferred from the image processor 204 to the image data control unit 203 is sent from the image data control unit 203 to the image memory access control unit 221 via the parallel bus 220.

Here, the system controller 23
1, control of access to image data and the memory module 222, expansion of print data of the external PC (personal computer) 223, and compression / expansion of image data for effective use of the memory module 222. .

The image data sent to the image memory access control unit 221 is stored in the memory module 222 after data compression, and the stored image data is read out as needed. The read image data is decompressed, restored to the original image data, and returned from the image memory access control unit 221 to the image data control unit 203 via the parallel bus 220.

After the image data is transferred from the image data control unit 203 to the image processor 204, image quality processing and video / video processing are performed.
The data control unit 205 performs pulse control, and the image forming unit 206 forms a reproduced image on transfer paper.

In the flow of the image data, the functions of the digital multi-function peripheral are realized by the bus control by the parallel bus 220 and the image data control unit 203. The facsimile transmission function performs image processing on the read image data by the image processor 204, and executes the image data control unit 2.
03 and the facsimile control unit 224 via the parallel bus 220. The facsimile control unit 224 performs data conversion to a communication network, and transmits the data to a public line (PN) 225 as facsimile data.

On the other hand, the received facsimile data is
Line data from the public line (PN) 225 is converted into image data by the facsimile control unit 224 and transferred to the image processor 204 via the parallel bus 220 and the image data control unit 203. In this case, no special image quality processing is performed, the dot rearrangement and the pulse control are performed in the video data control unit 205, and the reproduced image is formed on the transfer paper in the image forming unit 206.

In a situation where a plurality of jobs, for example, a copy function, a facsimile transmission / reception function, and a printer output function operate in parallel, the system controller assigns the right to use the reading unit 201, the imaging unit 206, and the parallel bus 220 to the job. 231 and the process controller 211.

The process controller 211 controls the flow of image data, and the system controller 231
Controls the entire system and manages the activation of each resource. In addition, the function selection of the digital multi-function peripheral is selectively inputted on the operation panel (operation unit) 234, and the processing contents such as the copy function and the facsimile function are set.

The system controller 231 and the process controller 211 communicate with each other via the parallel bus 220, the image data control unit 203, and the serial bus 210. Specifically, communication between the system controller 231 and the process controller 211 is performed by performing data format conversion for the data interface between the parallel bus 220 and the serial bus 210 in the image data control unit 203.

(Configuration of SIMD Processor) Next, the configuration of the SIMD processor will be described.
FIG. 3 is an explanatory diagram for explaining the configuration of the SIMD type processor. As shown in the figure, in this SIMD type processor, accumulators (A) corresponding to respective pixels are arranged in the main scanning direction.

Each accumulator has a 16-bit width A
In addition to the registers F and F, there are a plurality of 8-bit registers, which store calculation data and calculation results. Note that one unit corresponding to one pixel is 1
It is called a PE (processor element), and a plurality of PEs arranged in the main scanning direction that can be processed at one time is called one SIMD. In this SIMD type processor, each PE
Can refer to the values of the PEs next to the left and right (the values of the accumulator and the register).

(Configuration of Gradation Processing Unit) Next, image processing on the image processing unit 103 equipped with the SIMD type processor shown in FIG. 1 will be described. FIG. 4 is an explanatory diagram for explaining image processing on the image processing unit 103 equipped with the SIMD type processor shown in FIG.

The main process 401 shown in FIG. 6 selects each process according to the selected mode, and describes the selected process in a process call routine 402. Since each process is functionalized, the process call routine 402
The image processing functions 3 to 407 are called to sequentially process the input data.

At this time, processing such as feature amount extraction 404
An arbitrary feature is determined using the data after shading 403, and the result is stored in another register.
The input data itself is returned to the processing call routine without conversion, and is passed to the next filter processing 405.

The data input to the filter processing 405 is converted into appropriate correction data here, and the filter processing function returns the data after the filter correction to the processing call routine 402. The returned data is used in the next γ processing 4
06 is passed to the function. Thereafter, in the image processing 407 according to the present invention, the function performing the image processing 407 receives the data after the γ processing from the processing calling routine 402 and performs gradation processing, quantization, and density conversion.

FIG. 5 is an explanatory diagram for explaining the flow of each process shown in FIG. As shown in the figure, the scanner data 501 is first subjected to shading 502 and then branched to two processes. Specifically, while the image processing 506 is performed through the filter processing 504 and the γ correction 505, the feature quantity extraction processing 503 is performed, and this feature quantity is used in the image processing 506. Then, the processing result of this image processing 506 is output processing 507.

As described above, in this image processing apparatus, in addition to the image processing (gradation processing, quantization and density conversion) according to the present invention, shading 502 and filter processing 50 are performed.
4, γ correction processing 505, feature extraction 503, and the like.

(Concept of Image Processing) Next, the concept of image processing according to the present invention will be specifically described. FIG.
FIG. 3 is an explanatory diagram for explaining the concept of image processing according to the present invention. As shown in the figure, a function call (fu) corresponding to the processing calling routine shown in FIGS.
The actual processing function is called by (naction call) 601. Since the SIMD type processor is basically a sequence of processes, each process is sequentially performed even when separate processes as shown in the figure are performed and selection is performed according to the feature amount of an image. .

Specifically, first, the function call 601 calls a one-dot processing function, and performs one-dot multi-level (or binarization) processing 6 on the image data after γ correction.
Do 04. Here, a case in which one-dot processing is performed is shown, but the invention is not necessarily limited to one-dot processing.

Here, when performing the one-dot processing 604, vector data is created in addition to the quantization by referring to the contents of the register 603 storing the result of the feature amount extraction 602, and the quantization result and the vector data are stored. Is stored in the register 605. Since the image data input to the one-dot processing function is also used in the next gradation processing, the image data is returned to the function call 601 without any change.

Thereafter, the function call 601 is:
Call the dither function with vector, which is the gradation processing function,
The vector-added dither processing 606 is performed. Then, one of the processing result of the vector-added dither processing 606 and the one-dot processing result stored in the register 605 is selected based on the content of the register 603 storing the feature amount extraction result, and the selected result is used as a function. Returned to call 601.

(Dithering with Vector) Next, the dithering with vector 606 shown in FIG. 6 will be described more specifically. 7 and 8 are explanatory diagrams for explaining the vector-added dither processing 606 shown in FIG.

FIG. 8A is a diagram showing an example of a binary dither matrix. When dither processing is performed, the data in this matrix is used as a threshold, and the threshold and the image data are compared. Determines the output black and white according to the magnitude relationship.

P1 to p8 of D × D size shown in FIG.
Is one pixel at an input pixel density (for example, 600 dpi). Here, assuming that the output pixel density is 1200 dpi, a pixel to be output is divided into four pixels obtained by dividing D × D into four, and the pixel pattern is divided into 0 to gray scales shown in FIG.
There are four gradations. The square of each pattern shown in FIG.
This corresponds to one pixel of the output pixel density.

Here, considering a case where binarization is performed using the dither matrix shown in FIG. 8A and converted to the output pixel density shown in FIG. 7, how to fill four output pixels (from where to how) Can be indicated by a vector that is an arrow of size 1.

For this reason, as shown in FIG. 7, the vector as the arrow is associated with the dot formation, and the dither vector data shown in FIG. In this way, information on whether to fill the dots can be held.

(Procedure for Determining Threshold by Dither Processing) Next, a procedure for determining a threshold value by dither processing will be described. FIG. 9 is a flowchart showing a procedure for determining a threshold by dither processing. Here, it is assumed that the output pixel density is twice the input pixel density and the output value is binary.

As shown in the figure, first, the threshold value of the dither matrix corresponding to each position information is set to each P.
The data is loaded into the register of E (step S901).
Then, when the input data is completed for one SIMD and the threshold value is stored in each register, the PEs for one SIMD compare the input data and the threshold value all at once (step S902).

If the result is less than the threshold value (No at Step S902), a flag 1 is set to 0 (Step S910), and if the result is larger than the threshold value (Yes at Step S902), the comparison is made. The result is set to 1 (step S903). Thereafter, a line delay is performed (step S904). When the output results for the three lines are accumulated, each of the PEs on the middle line obtains the sum of the four adjacent pixels located above, below, left and right of the PE (step S90).
5) The sum is used to determine whether the number of gradations is one, two, three, or four (step S906).
909).

For example, if the input data is equal to or larger than the threshold value, 1 is set, the binary result of the left and lower PEs is 1, and the other 2
If the binarization result is 0, the sum is 2, so that two gradations are set. Then, the actual dot formation is determined by combining the vector data shown in FIG. 8B with the vector data. If the arrow is vector data such as p1 from the top to the bottom, a 4-pixel pattern in which the upper two are 1 and the lower two are 0 as represented by the two gradations of p1 is output.

FIG. 10 is a schematic diagram schematically showing an example of SIMD movement. 1001, 1002, 1 shown in FIG.
Reference numeral 003 denotes one pixel having the input pixel density. Here, assuming that the output pixel density is doubled, the control data 10
05 to p1 to p6 for each PE
To enter.

Then, the threshold values corresponding to the respective pixel data are stored in the registers r11 to r16, a comparison operation is performed when the data is completed, and 1 is stored in A1 to A6 of the PE where no carry occurs. A1 of the PE with carry
0 is stored in .about.A6. And the result is FIFO
Using 1006 to 1008, three lines are stored.

When attention is paid to r33, the binarized data r33 is 1, and output data is generated. Therefore, in order to determine the number of gradations, the upper and lower (r23, r43) and the two adjacent (r31, r43) are determined. r35) is obtained. Here, since the sum is 2, two bits 1 are set in A3. If the sum is 3, 11 is set for A3 and 10 is set for A4. If the sum is 4, A3 and A4 are set.
Set 11 for both.

Note that these two bits are represented by r
When stored in 53 and r63, output pixel data obtained by directly performing double density conversion is obtained. Since A4 is 00, the output pixel data of r53 and r63 both become 0. The pixel image after the density conversion is indicated by a line 101.
1, 1012, and the input data of 1002 is converted as d1 to d4.

Further, since an SIMD type processor is used here, these operations can be performed simultaneously for one SIMD. In particular, since the same data is stored twice in the PE here, for example, even-numbered PEs are stored.
The same processing can be performed all at once.

(Multi-Valued Dither Matrix) Next, a case where the number of gradations is determined using the multi-valued dither matrix will be described. FIG. 11 is an explanatory diagram for explaining a case where the number of gradations is determined using the multi-value dither matrix. It is assumed that the correspondence between the vector data and the dot formation pattern is as shown in FIG.

Assuming that there are four dither matrices shown in FIG. 9A and vector data shown in FIG. 9B, the number of gradations can be obtained by sequentially comparing the four dither matrices. it can. In other words, a result equivalent to the conventional method of obtaining a density-converted image after multi-level conversion is once obtained, but since dither has vector data in addition to the threshold value, a flag is set according to the vector data. By doing so, it is possible to directly obtain a binary dot pattern after density conversion without multi-value conversion.

FIG. 12 is a flowchart showing a processing procedure for determining a gradation in accordance with the vector data shown in FIG. As shown in the figure, since there are four dither matrices A to D shown in FIG. 11A, first, it is checked whether the data is smaller than the threshold value A of the dither matrix A. If it is smaller (Yes at Step S1201), the gradation is set to 0 without flag (Step S1202).

On the other hand, if the data is equal to or larger than the threshold A (No at Step S1201), after setting the flag 1 (Step S1203), it is determined whether the data is smaller than the threshold B of the dither matrix B or not. (Step S12)
04) If it is small (Yes at Step S1204), the gradation is set to 1.

On the other hand, if it is equal to or larger than the threshold value B (No at Step S1204), the flag 2 is set (Step S1).
205), it is confirmed whether or not the data is smaller than the threshold value C of the dither matrix C (step S1206). If it is smaller (Yes at step S1206), the gradation is set to 2.

On the other hand, if the data is equal to or larger than the threshold C (No at Step S1206), after setting the flag 3 (Step S1207), it is determined whether or not the data is smaller than the threshold D of the dither matrix D. (Step S12)
08), if it is smaller (Yes at Step S1208), the gradation is set to 3, and if it is equal to or larger than the threshold value D (Step S12).
08 No), the tone is set to 4 after setting the flag 4 (step S1209).

(Relationship between Image Features and Vector Data)
Next, the relationship between the feature amount of an image and vector data will be described. FIG. 13 is an explanatory diagram for explaining the mutual relationship between the feature amount of an image and vector data. As shown in the figure, vector data 1302a to 1302d correspond to oblique edge detection in feature amount extraction 1301, vector data 1303a to 1303d correspond to black character detection,
Vector data 1304a to 1304 are used for non-dot detection.
d corresponds.

That is, a thin oblique line or the like can be smoothly reproduced by associating oblique vector data with an image portion having an oblique edge portion. Further, when the vector data of the dither generates a dot pattern that generates a halftone dot, the character or the like becomes difficult to read due to the dither halftone dot or the like. Therefore, characters and the like can be made easier to read by associating the vector data with the non-halftone portion so that the halftone generation is broken and the vertical line base tone is obtained.

Further, when the gradation processing is switched in accordance with the feature amount of the image, the vector data to be provided for the one-dot processing, the error diffusion processing, and the like can be based on vertical lines such as 1304a to 1304d. Furthermore, in the case of a color original, if a character is determined as a black character, a character-oriented dot pattern is used to clarify the character, and otherwise, a dithered dot pattern with stable dots is used to obtain a high-gradation photographic image. Can be.

(Another Concept of Image Processing) By the way, in FIG. 6 and the description thereof, the concept of the image processing in the case where the separate processing is performed and the selection is performed according to the feature amount of the image has been described. The vector data can be adjusted later. Therefore, the concept of image processing in this case will be described.

FIG. 14 is an explanatory diagram for explaining the concept of image processing in the case of adjusting vector data after dither processing with a vector. As shown in the figure, a function call 1401 is:
First, a feature amount extraction 1402 is called to extract, for example, image feature amounts as shown in FIG. 13 from shading-corrected image data.

Thereafter, vector data generation 1403 is performed using the extraction result, and the generated vector data for one SIMD is stored in the register 1404. In addition,
The image data passed to the feature extraction 1402 is returned to the function call 1401 without being changed.

Then, the function call 1401
Invokes the vector-added dither processing 1405 at the gradation processing stage, quantizes the image data after the γ correction, and obtains vector data. After that, referring to the vector data stored in the register 1404, for the part where the vector data based on the feature amount is created, the vector data adjustment 1406 that replaces the vector data of the dither with the vector data stored in the register 1404 is performed. The result is output as data (outdata) 1407 as a pixel pattern of output pixel density.

(Example of a ternary dot pattern)
An example of a ternary dot pattern with a final output result will be described. FIG. 15 is an explanatory diagram illustrating an example of a ternary dot pattern with a final output result. The white squares shown in the figure are data 0 (WHITE), the halftone squares are data 1 (GRAY), and the black squares are data 2 (BLA).
CK).

For example, for D1a, one of the four is a black square (BLACK), which is 2
The gradation is equivalent to the case where one output pixel is a halftone dot (GRAY). Also, two black squares of D1c,
The gradation becomes equivalent to four squares of halftone dots of D1d.

When forming dots in an oblique pattern, two pixels can be reduced to half the density as in D5b, instead of forming one as a black square as in D5a. Also, by replacing D6c with D6d or D6e, jaggies at the edges can be eliminated and a smooth oblique edge can be obtained.

As described above, when the quantized value to be output is three or more, the smoothness of the edge portion can be improved by using the dot pattern using the gray (GRAY) data according to the characteristic amount. it can. In the dot patterns D1d to D4d shown in the figure, since four pixels can be simultaneously filled with halftone dots (GRAY), noise (low-density pixels causing white spots) in solid black is easily filled. It is used to prevent white spots and obtain a good image.

(Example of Image Improvement) Next, an example of image improvement using a ternary dot pattern will be described. FIG. 16 is an explanatory diagram for describing an example of image improvement using a ternary dot pattern. FIG. 1A shows an example of image data based on an input pixel density, and FIG. 1B shows image data to which a dot pattern A and a dot pattern B are applied when converting to an output pixel density. I have. As can be seen from FIG. 7B, if the dot pattern B is used when converting to the output pixel density, a smooth edge portion can be obtained.

FIG. 10C shows an example of image data based on the output pixel density. When the hatched portions of this image data are replaced with dot patterns D1 and D2, the image data shown in FIG. Is obtained, but this image data cannot be said to be a smooth image. On the other hand, the dot pattern C1 and the dot pattern C
When 2 is used, smooth image data without solid omission as shown in FIG.

Next, a processing procedure for obtaining a ternary output pixel using the four dither matrices shown in FIG. 11C will be described. FIG. 17 is a flowchart showing a processing procedure for obtaining a ternary output pixel using the four dither matrices shown in FIG. 11C.

As shown in the figure, first, it is determined whether or not the input data is smaller than a threshold value TH1 (step S1701). If the input data is smaller than the threshold value TH1 (step S1701 affirmation), Number of gradations (dout) = 0
(Step S1709).

On the other hand, when the input data is equal to or larger than the threshold value TH1 (No at step S1701), the input data is set to the threshold value TH2.
It is determined whether the value is smaller than the threshold value TH2 (step S1702).
2 affirmative), the number of gradations (dout) = 1 (step S
1708).

If the input data is equal to or larger than the threshold value TH2 (No at Step S1702), the input data is set to the threshold value TH3.
It is determined whether or not the threshold value is smaller than the threshold value TH3 (step S1703).
3 is affirmative), the number of gradations (dout) is set to 2A (step S1707).

If the threshold value is equal to or greater than the threshold value TH3 (No at step S1703), the input data
It is determined whether it is smaller than H4 (step S170).
4) If it is smaller than the threshold value TH4 (Yes at Step S1704), the number of gradations (dout) is set to 3 (Step S1706). If it is equal to or more than the threshold value TH4 (No at Step S1704), the gradation is set. The number (dout) is set to 4 (step S1705).

As described above, when the input data is compared with each threshold value and the number of gradations of the output data is determined from the result, the output pixel data is determined based on the number of gradations. . For example, the number of gradations 2 (dout) = 2A
In the case of, the presence / absence of feature vector data is referred to the PE of interest. If the feature vector data exists, the output data is replaced with the vector pattern B. If the feature vector data does not exist, the vector is output. Pattern A can be used as output data. The magnitude of the vector of the vector pattern B is 0.

[0109]

As described above, according to the first aspect of the present invention, the characteristics of an image are determined from the input image data, and the determination result and the pixel value of each pixel forming the image data are determined. The first vector data of the output pixel is determined based on the determined result, and either the second vector data or the first vector data stored in the dither table is selected based on the determination result. Since the gradation processing is performed based on the stored threshold value, even if the number of gradations to be converted is a small value, it is possible to obtain a good density-converted image without image deterioration such as data omission. In addition, an image processing apparatus capable of performing high-speed density conversion utilizing the advantage of the SIMD type processor can be obtained.

According to the second aspect of the present invention, the first vector data determined by the determining means is selected with priority over the second vector data stored in the dither table. In addition, it is possible to obtain an image processing apparatus that can easily adjust the dot arrangement and convert the dot arrangement in accordance with the characteristics of the image and obtain a high-quality density-converted image.

According to the third aspect of the present invention, when the quantization value is 3 or more, the vector data having the vector size of 0 can be selected.
An effect is obtained in that an image processing apparatus capable of easily performing a dot arrangement conversion in which an edge portion is smooth or a dot conversion in which solid filling is good in a small number of steps and thereby obtaining a good density conversion image is obtained.

According to the fourth aspect of the present invention, the characteristics of the image are determined from the input image data, and the output pixels are determined based on the determination result and the pixel value of each pixel forming the image data. 1 is determined, and either the second vector data or the first vector data stored in the dither table is selected based on the determination result, and the selected vector data and the threshold value stored in the dither table are selected. Since the gradation processing is performed on the basis of the above, even if the number of gradations to be converted is a small value, it is possible to obtain a good image after density conversion without image deterioration such as data drop, and the like. An effect is obtained that an image processing method capable of performing high-speed density conversion utilizing the superiority of the type processor can be obtained.

According to the fifth aspect of the present invention, the first vector data determined by the determining means is selected with priority over the second vector data stored in the dither table. In addition, there is an effect that an image processing method capable of easily performing adjustment of the dot arrangement and conversion of the dot arrangement according to the characteristics of the image and obtaining a high-quality density-converted image is obtained.

According to the sixth aspect of the present invention, when the quantization value is 3 or more, the vector data whose vector size becomes 0 can be selected.
An effect is obtained that an image processing method capable of easily performing a dot arrangement conversion in which an edge portion becomes smooth or a dot conversion in which solid filling is good in a small number of steps and thereby obtaining a good density conversion image is obtained.

According to the seventh aspect of the present invention, a program for causing a computer to execute the method according to any one of the fourth to sixth aspects is recorded, so that the program becomes machine-readable. Thus, there is an effect that a recording medium capable of realizing the operations of claims 4 to 6 by a computer can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram functionally showing a configuration of an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating an example of a hardware configuration of the image processing apparatus according to the embodiment;

FIG. 3 is an explanatory diagram for describing a configuration of a SIMD type processor.

FIG. 4 shows a SIMD (Single Ins) shown in FIG.
fraction Multiple Datast
FIG. 4 is an explanatory diagram for describing image processing on an image processing unit equipped with a (ream) type processor.

FIG. 5 is an explanatory diagram illustrating a flow of each process illustrated in FIG. 4;

FIG. 6 is an explanatory diagram for explaining the concept of image processing according to the present invention.

FIG. 7 is an explanatory diagram for explaining the dither processing with a vector shown in FIG. 6;

FIG. 8 is an explanatory diagram for explaining the vector-added dither processing shown in FIG. 6;

FIG. 9 is a flowchart illustrating a threshold determination procedure by dither processing.

FIG. 10 is a schematic diagram schematically illustrating an example of SIMD movement.

FIG. 11 is an explanatory diagram for explaining a case where the number of gradations is determined using a multi-value dither matrix.

FIG. 12 is a flowchart showing a processing procedure for determining a gray scale according to the vector data shown in FIG. 11;

FIG. 13 is an explanatory diagram for explaining a mutual relationship between a feature amount of an image and vector data.

FIG. 14 is an explanatory diagram for explaining the concept of image processing when adjusting vector data after dithering with a vector;

FIG. 15 is an explanatory diagram showing an example of a ternary dot pattern with a final output result.

FIG. 16 is an explanatory diagram for describing an example of image improvement using a ternary dot pattern.

FIG. 17 is a flowchart showing a processing procedure for obtaining a ternary output pixel using the four dither matrices shown in FIG. 11C.

[Explanation of symbols]

 REFERENCE SIGNS LIST 100 image data control unit 101 image reading unit 102 image memory control unit 103 image processing unit 104 image writing unit 201 reading unit 202 sensor board unit 203 image data control unit 204 image processing processor 205 video data control unit 206 image formation Unit (engine) 210 Serial bus 211 Process controller 212,232 RAM 213,233 ROM 220 Parallel bus 221 Image memory access control unit 222 Memory module 223 Personal computer (PC) 224 Facsimile control unit 225 Public line 231 System Controller 234 Operation panel 401 Main processing 402 Processing call routine 403 Shading 404 Feature extraction 405 Filter processing 406 γ processing 407 Image processing

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04N 1/40 101E (72) Inventor Yasuyuki Nomizu 1-3-6 Nakamagome, Ota-ku, Tokyo RICOH Company, Limited (72) Inventor Hideto Miyazaki 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd. (72) Inventor Yoshiyuki 1-3-6 Nakamagome, Ota-ku, Tokyo Ricoh Co., Ltd. (72 Inventor Yuji Takahashi 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Company (72) Inventor Hiroyuki Kawamoto 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Company (72) Inventor Wood Sugitaka Ricoh, 1-3-6 Nakamagome, Ota-ku, Tokyo (72) Inventor Shinya Miyazaki 1-3-6 Nakamagome, Ota-ku, Tokyo (72) Inventor Takusho Fukuda 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd. (72) Inventor Takeharu Tone 1-3-6 Nakamagome, Ota-ku, Tokyo Ricoh Co., Ltd. F Term (reference) 5B045 AA01 GG14 5B057 BA02 CD06 CE06 CE13 CH04 CH07 DA20 DC16 DC22 DC25 5C076 AA21 BA06 BA07 BB04 5C077 LL19 MM03 MP01 MP08 NN06 NN14 NN15 PP01 PQ12 PQ23 RR19

Claims (7)

[Claims] An input image data read from an image reading unit is set to a SIMD so as to conform to an output density of an image output unit.
(Single Instruction Multi
An image processing apparatus for performing density conversion using a ple Data stream type processor, a determination unit for determining characteristics of an image from input image data, a determination result by the determination unit, and a pixel of each pixel forming the image data Determining means for determining the first vector data of the output pixel based on the value; a dither table storing a predetermined threshold value and second vector information; and a second vector data stored in the dither table or Selecting means for selecting any of the first vector data determined by the determining means based on the determination result by the determining means; and selecting the vector data selected by the selecting means and a threshold value stored in the dither table. And a gradation processing means for performing gradation processing based on the image. Management apparatus. 2. The apparatus according to claim 1, wherein the selection unit selects the first vector data determined by the determination unit prior to the second vector data stored in the dither table. The image processing apparatus according to any one of the preceding claims. 3. The image processing apparatus according to claim 1, wherein said gradation processing means can select vector data having a vector size of 0 when the quantization value is 3 or more.
An image processing apparatus according to claim 1. 4. An image processing method for converting the density of input image data read from an image reading unit using a SIMD type processor so as to conform to the output density of an image output unit. A determination step of determining; a determination step of determining first vector data of an output pixel based on a determination result of the determination step and a pixel value of each pixel forming the image data; A selection step of selecting either the vector data or the first vector data determined in the determination step based on a determination result in the determination step; and storing the vector data selected in the selection step and the dither table. A gradation processing step of performing gradation processing based on a threshold value; Law. 5. The method according to claim 4, wherein in the selecting step, the first vector data determined in the determining step is preferentially selected over the second vector data stored in the dither table. The image processing method described in the above. 6. The gradation processing step according to claim 4, wherein, when the quantization value is 3 or more, vector data having a vector magnitude of 0 can be selected.
The image processing method according to 1. 7. A computer-readable recording medium on which a program for causing a computer to execute the method according to claim 4 is recorded.