JP2008141502A - Image processing method, image processing device, and system thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable high-speed color conversion processing to compressed image data by performing cache processing by a multithread. <P>SOLUTION: An image processing method prepares cache data for color conversion by multithreads 344-347 according to a DC component 341 per processing unit block of image data when image data 316 compressed is decompressed by a decompression processor 320, and stores the data in a global range 348. When image data 331 decompressed is processed by a color conversion processor 351, the cache data within the global range 348 is referred. Thus, high-speed color conversion processing to the compressed image data is realized. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image processing method, an image processing apparatus, and a system thereof, and more particularly to an image processing method, an image processing apparatus, and a system thereof that perform color matching on image data.

  Many methods have been proposed in which an image is mapped in the frequency domain, and quantization and coding suitable for each frequency component are performed to realize efficient image compression. In particular, an image adaptive image compression method combining a discrete cosine transform (DCT) and a Huffman code is widely used, such as being adopted in an international standard. These image compression techniques are often used to compress high-definition (high-resolution) color images with the recent spread of color image systems.

  In general, as a color image system, a color printing system using a color ink jet printer is in widespread use. In such a color image system, a high-definition color image is compressed and stored in a disk device (file system), the compressed data is expanded as necessary, and color printing is frequently performed. Yes.

  Here, an example of a general image processing system using an image compression technique will be outlined below with reference to FIG. In FIG. 11, original image data 111 that is color image data is compressed by a compression processing unit 110. In the compression processing unit 110, the input original image data 111 is first converted into frequency data in units of 8 × 8 pixel blocks for each color component plane by the DCT conversion processing unit 112, and for each frequency component by the quantization processing unit 113. Quantized as appropriate. Next, the DC coefficient that is the DC component of the frequency data is calculated by the DC coefficient processing unit 114 as a difference from the preceding and following blocks, and the AC coefficient that is the AC component of the frequency data is entropy encoded by the AC coefficient processing unit 115. The These encoded data groups constitute compressed data 116.

  The compressed data 116 obtained from the compression processing unit 110 is decompressed and output as necessary by the decompression processing unit 120. In the decompression processing unit 120, first, the inverse AC coefficient processing unit 121 performs entropy decoding to obtain a quantized AC coefficient, and the inverse DC coefficient processing unit 122 obtains the DC coefficient of the block of interest from the preceding and succeeding blocks. From the quantized data group thus obtained, the inverse quantization processing unit 123 and the inverse DCT transform processing unit 124 obtain decompressed data 131 that is color image data.

  The obtained decompressed data 131 is mapped into a color space suitable for printing (color matching: RGB → YMCK) by the color conversion processing unit 141. The image data after the color conversion is further enlarged (or reduced) by the resolution conversion unit 151 and adapted to the printing resolution to form an image output signal 161.

  In the image processing system shown in FIG. 11, since the data size of the decompressed data 131 becomes large, there is a problem that the color matching processing time in the color conversion processing unit 141 becomes long. In order to perform the color matching process at high speed here, a “caching process” or the like that temporarily holds a color conversion result that has been converted has been proposed (for example, see Patent Document 1).

  The processing speed in the color printing system described above greatly depends on the size of image data to be processed. For example, when processing a high-resolution photographic image or the like, the processing speed tends to decrease.

  In recent years, the resolution of digital cameras has been remarkably increased, and it is not uncommon for a high-end single-lens reflex digital camera to have more than 10 megapixels. These high-end machines are equipped with a large CMOS sensor, which enables generation of an image with excellent resolution. For example, a model that generates an image having approximately 16.7 million effective pixels (16 megapixels) has appeared.

  An image taken with a 16-megapixel digital camera consists of 4992 × 3328 pixels (16.61 million pixels). In general, since the photographed photographic image data is an RGB image, a 47.53 MB memory is required to develop it on the host PC.

  Furthermore, when printing an RGB image photographed with a digital camera, it is necessary to convert it into a color space such as CMYK. For this reason, a printing system that prints out a photographed image must be equipped with a color management system (CMS). For example, a color LUT is used to perform color matching processing that matches the characteristics of the output device. Printed.

FIG. 12 shows a schematic configuration of a conventional color printing system for printing a photographed image. In the figure, the host PC 100 and the printer controller 200 are connected by a serial cable or a network cable. In this system, since the bandwidth for data transfer in the connectivity (transmission path) is narrow, the print job is configured with a reduced data size. Therefore, when printing a compressed image, the host PC 100 does not perform the compressed image expansion process, and transfers the data to the printer controller 200 in a compressed state. On the printer controller 200 side, the RIP manager decompresses the received compressed image, performs color matching, and prints it out.
JP 2000-113172 A

  In the above-described conventional color image system, it has been described that the caching process is performed for the purpose of speeding up the color matching process. However, this caching process is particularly effective for vector graphics and the like, but has not been particularly effective when processing compressed image images and the like.

  Further, with the development of current communication technology, it is also possible to sufficiently widen the bandwidth in connectivity between the host PC and peripheral devices even in the configuration shown in FIG. In such a printing environment, it is not always necessary to reduce the print data transmission size. In addition, since the processing speed on the host PC side is also dramatically improved, it is considered that the processing speed is often improved overall when data processing such as decompression is performed on the host PC side.

  For example, in the case where the applicant experimentally performs a color matching process with linear 8-point interpolation on an 8-bit image of 4992 × 3328 pixels (about 48 MB) on a general high-performance host PC, about 6.5 seconds. As a result, the process was completed.

  However, in current commercial printers, it is necessary to complete processing within 2 seconds to 1 second per image in order to enable high-speed printing of 30 to 60 sheets or more per minute. Therefore, even if the color matching processing in the color printing system as shown in FIG. 12 can be realized on the host PC side, if the processing time is still in the above range, such high-speed printing cannot be handled.

  The present invention has been made to solve the above-described problems individually or collectively, and an object thereof is to provide an image processing method, an image processing apparatus, and a system for realizing high-speed color matching processing.

  As a technique for achieving the above object, the image processing method of the present invention includes the following steps.

  That is, an expansion step for expanding compressed image data, a cache generation step for generating cache data for color conversion based on frequency components for each processing unit block of the image data, and for the expanded image data And a color conversion step for performing a color conversion process based on the cache data.

  An image processing method for performing color conversion processing on image data, wherein a test step for experimentally performing color conversion processing with a plurality of block sizes on the image data, and a test result in the test step A block size specifying step for specifying a block size for which the color conversion process has been executed in the shortest time; and a size information adding step for adding size information indicating the block size to the image data. The color conversion process is performed with a block size indicated by the size information.

  The image data further includes an interpolation calculation step for executing an interpolation calculation for color conversion for a target color of the color conversion process, and a result holding step for storing the calculation result in the interpolation calculation step in a file. A color conversion process is performed on the basis of the file.

  An image processing method for performing color conversion processing on image data in a host computer that enables parallel processing by a plurality of threads, the test step for performing the color conversion processing on a trial basis with a variable number of threads, and the test Based on the test results in the step, the number of threads specifying step for specifying the number of threads that executed the color conversion processing in the shortest time, and the number of threads information that holds thread number information indicating the number of threads in the system file in the host computer And a color conversion process for the image data is performed as a parallel process based on the number of threads indicated by the thread number information.

  With the above configuration, according to the present invention, high-speed color matching processing is realized.

  Hereinafter, the present invention will be described in detail based on preferred embodiments with reference to the accompanying drawings. The configurations shown in the following embodiments are merely examples, and the present invention is not limited to the illustrated configurations.

<First Embodiment>
First, an overview of image processing in the printing system of the present embodiment will be described.

  FIG. 1 shows a schematic configuration example of an image forming unit in the color printer 2010 of the present embodiment. In the figure, reference numerals 2200 to 2230 denote color ink cartridges filled with yellow (Y) color ink, magenta (M) color ink, cyan (C) color ink, and black (K) color ink, respectively. Each of the ink cartridges 2200 to 2230 and the pressure pump 2160 are connected by independent pipes for supplying ink, and ink of each color is pumped from the pressure pump 2160 to the print head 2120 at a constant pressure. ing.

  The recording paper is supplied from the rear surface of the color printer 2010 and is fixed by a platen 2110 and a front guide roller 2150. The operation of the color printer 2010 is performed by pressing a key on the control panel 2180, and the control board 2190 controls all operations of the color printer 2010.

  In order to perform printing by the color printer 2010, a print control command and print data may be sent to the control board 2190 via an interface (not shown). For example, when RGB is designated in the color designation command, the control board 2190 that has received the print control command and the print data drives the print head 2120 after converting the RGB data into CMYK data by the internal color processing unit. Print.

  In the printing system of this embodiment, compressed color image data (hereinafter simply referred to as compressed data) is restored to the original color image data by the decompression process, and at the same time, frequency component data is extracted. . Since the compressed data is divided into a plurality of data blocks (usually rectangular areas), for example, only the DC component is extracted for each block area from one image data, and only the color is preceded. Color matching processing can be performed. Then, in order to reuse the color matching result in subsequent processing, the data processing speed as a whole can be improved by temporarily holding (caching) the result.

FIG. 2 is a diagram showing an outline of image processing in the printing system of the present embodiment. In the figure, for example, original image data 311 having 8 bits for each color component is compressed by a compression processing unit 310. In the compression processing unit 310, the input original image data 311 is first converted into frequency data by the DCT conversion processing unit 312 for each color component plane, for example, in units of 8 × 8 pixel blocks. Here, if the function indicating the original image is i (x, y) and the corresponding DCT function is I (u, v), the conversion equation is defined as follows.
I (u, v) = 2c (u) / N ・ 2c (v) / N
・ ΣΣi (x, y) ・ cos {(2x + 1) uπ / 2N} ・ cos {(2y + 1) vπ / 2N} ... (1)
However, 0 ≦ u, v ≦ N-1
When u, v = 0, c (u) = c (v) = 1 / √2
When u, v = 1,2, ..., N-1, c (u) = c (v) = 1
Note that the range of the first Σ operation in equation (1) is x = 0 to N,
The range of the second Σ operation is y = 0 to N.

  Then, the converted data is quantized by the quantization processing unit 313 for each frequency component. This quantization includes processing such as scaling of the number of bits. Then, the DC coefficient that is the direct current component of the frequency data is calculated by the DC coefficient processing unit 314 to calculate the difference from the preceding and following blocks, and the AC coefficient that is the alternating current component is entropy encoded by the AC coefficient processing unit 315, Compressed data 319 is configured.

  The compressed data 316 obtained from the compression processing unit 310 is temporarily stored in a storage device (not shown), and is decompressed and output as necessary. Hereinafter, the decompression process in the present embodiment will be described in detail.

  The compressed data 316 input to the decompression processing unit 320 is first entropy-decoded by the inverse AC coefficient processing unit 321 to obtain quantized AC coefficients, and the inverse DC coefficient processing unit 322 calculates the DC coefficient of the target block from the preceding and following blocks. Is required. These quantized data groups are inversely quantized (and inversely scaled) by the inverse quantization processing unit 323. The frequency component for each region obtained at this time is stored in a storage device (not shown) as an image frequency component (DCT coefficient data) 341 or temporarily output to a memory.

  The image frequency component (DCT coefficient data) 341 is determined by the candidate color selection unit 342 as to whether or not the color constituting the DC component of the block is a color to be cached. Specifically, the candidate color selection unit 342 checks the image frequency component 341, and the cache thread is activated only when the low frequency component is larger than a preset threshold. Here, the term “cache thread” refers to performing color matching processing, and the color matching result is temporarily held in the global area (cache table) 348 so that it can be reused in subsequent processing.

  Reference numeral 343 denotes a cache thread activation unit that activates a cache thread. In the present embodiment, up to four color conversion thread processes 344 to 347 can be activated. Since the entry to each thread goes through the queue queue, if all four threads have already been started, the thread enters the queue as appropriate, and waits until any processing becomes available. As described above, the result of the color matching processing (for example, sRGB → printer CMYK) as the thread processing is stored in the global area 348.

  Here, FIG. 3 shows a configuration example of image data processed in this embodiment. As shown in the figure, the image data in this embodiment is composed of an 8 × 8 pixel rectangular area, that is, a block. According to this embodiment, for each block of image data, the frequency component (DCT coefficient data) 341 is determined by the candidate color selection unit 342, and if the color is to be cached, the cache thread activation unit 343 performs multi-cache processing. Start a thread.

  As described above, according to the present embodiment, cache processing by multi-thread is performed on one piece of image data composed of a plurality of blocks (Block 0 to n), so that cache data can be created at high speed. .

  Returning to the description of FIG. 2, the frequency data group obtained by the above-described inverse quantization processing unit 323 is subsequently mapped to the spatial domain by the inverse DCT transform processing unit 324. Since this process can be executed independently for each block, a plurality of threads are activated and the process is advanced in parallel. By the series of processes described above, decompressed data 331 that is color image data is obtained.

  The decompressed data 331 obtained in this way is converted into a device color space by the color conversion processing unit 351. In the present embodiment, at the time of this color conversion processing, the processing speed can be improved by performing the color matching processing while appropriately referring to the cache data stored in the global area 348.

  After the color conversion in the color conversion processing unit 351 described above, the resolution conversion processing unit 361 performs an enlargement (or reduction) process so as to match the print resolution, for example, and output image data 371 is obtained.

  In this embodiment, the image block is not limited to a rectangle. For example, the image blocks may be substantially hexagonal, each block may have an irregular size and shape, or there may be overlap between adjacent blocks.

  Further, the color conversion processing method in the color conversion processing unit 351 is not particularly limited as long as a high-quality image can be formed according to the frequency component of the image.

  Further, in the decompression processing unit 320 that decompresses the encoded image data, it is not always necessary to accompany the reproduction of the original image, and only necessary information is obtained from the encoded data. You may comprise so that a process may be performed. For example, the image data may be decoded into an intermediate code without being reproduced. That is, the decompression processing unit 320 in the present embodiment may be a unit that generates the original image partially or temporarily.

  As described above, according to the present embodiment, when the compressed image data is decompressed and output, the cache data is efficiently generated by the multi-thread, and therefore, color conversion processing referring to this is performed. As a result, the processing time of the entire color matching process is shortened.

<Second Embodiment>
Hereinafter, a second embodiment according to the present invention will be described. In the second embodiment, high-speed color matching is realized in a printing system that prints out a high-resolution image captured by a high-end digital camera.

  FIG. 4 is a diagram illustrating a schematic configuration of a printing system according to the second embodiment. In this system, when the connectivity (transmission path) of data transfer from the host PC 100 to the printer controller 200 is thin, for example, the connection between the host and the printer is connected by a remote network or the like, and it is difficult to secure a high-bandwidth transmission path. Assume the case.

  The color matching processing in the system shown in FIG. 4 is realized by software processing using a CPU inside the printer controller 200. The CPU used inside the printer controller 200 operates only with a single thread, but is equipped with a cache for shortening the data loading time and realizing high-speed processing.

  FIG. 5 is a diagram showing an outline of a cache built in the CPU in the printer controller 200 of the second embodiment. The CPU 220 shown in the figure includes a high-speed SRAM 222 as a cache memory. As a method for selecting an internal line for a cache, many devices for speeding up have been proposed, such as a full associative method, a direct map, and a 4-way set associative method.

  The main memory 231 that is an external memory is a large-capacity memory such as DDR, and an area of about 80 MB or more is allocated in order to realize color image processing in the second embodiment.

  Hereinafter, the speeding up of the color matching process in the second embodiment will be described in detail. In the second embodiment, the CPU 220 implements a color matching process by executing a software program (hereinafter simply referred to as a program). In order to improve the execution speed, the following (1), (2) The following control is performed.

  (1) In order to reduce the processing time required for the interpolation calculation associated with color matching, the calculation is performed in advance and the calculation result is stored in the main memory 231 as an LUT (lookup table).

  This LUT is a direct correspondence between sRGB and printer CMYK values, and is loaded into the DRAM 224, which is an internal memory where color matching processing starts. In order for the LUT to have all the calculation results, it is necessary to store a storage area of 64 MB, that is, the number of colors to the cube of 256 colors that can be expressed in 8 bits, that is, 64 MB. If the LUT capacity is large and the entire LUT cannot be stored in the DRAM 224, the portion of the LUT that is used for color matching processing is adaptively loaded.

  (2) Block the image so that the cache is effective. At this time, the effective block size is investigated in advance on the host PC side.

  The control shown in the above (1) and (2) will be described in more detail.

  First, the reduction in interpolation calculation time using the LUT of (1) will be described. When an execution time is analyzed with a performance analyzer or the like for an ordinary program that executes an interpolation operation associated with color matching, it can be seen that a considerable number of clocks are consumed for the multiplication / subtraction operation necessary for the interpolation operation. This is the same even if the processing order of the calculation formula for the interpolation calculation is switched. Therefore, in the second embodiment, the basic calculation formula of the interpolation calculation is left as it is, and the processing time is reduced by using a previously created LUT.

  That is, instead of starting an interpolation calculation associated with color matching after a job is submitted, the interpolation calculation is performed in advance for all colors in the input color space color, and the calculation result is stored in the printer controller 200. Stored in an LUT file or the like. The creation of the LUT file is not necessarily performed by the printer controller 200. If the created LUT is held in a place where high-speed access is possible in the printer controller 200, for example, the creation is performed on the host PC 100 side. You may do it. Here, the size of the LUT is 64 MB in total for 16.67 million colors when the input color space color is RGB 8-bit.

  At the time of color matching, the LUT file is loaded into the internal memory (DRAM 224), and the color matching value is performed at high speed by accessing the memory and using the LUT to process the input job. .

  In the second embodiment, in addition to the use of the interpolation operation LUT in (1) above, the cache in (2) is used more effectively, that is, the cache hit rate is improved, thereby further increasing the speed.

  Hereinafter, the improvement of the cache hit rate by the blocking of (2) will be described. In the second embodiment, the cache hit rate is improved by switching the loop unit in the CMS as described below.

  In general, it is considered difficult to capture the characteristics of the distribution of pixel values in an image. However, the local area of a block in the image has a certain degree of similarity (probabilistically) among the constituent pixels, which is sufficient to realize high speed processing by effectively using the cache. Furthermore, the hit rate of the data cache in the CPU can be improved by appropriately setting the block size as a CMS loop unit. Here, since the optimum block size for improving the cache hit rate varies depending on the image, the actually effective block size needs to be inspected for each original image.

  Accordingly, in the second embodiment, the block size inspection unit 130 in FIG. 4 performs, for example, four types of block sizes of 16 × 16, 32 × 32, 64 × 64, and 128 × 128 on the host PC 100 side. Check speed. Of course, the block size to be inspected here is not limited to these four types, and may be five types including 8 × 8, for example. Then, the optimum block size obtained by the inspection is pasted on the tag portion of the image file to be processed. FIG. 4 shows an example of an image file in which the optimum block size information 1401 is pasted as a tag.

  Here, the inspection procedure in the block size inspection unit 130 will be described with reference to the flowchart shown in FIG.

  When a program for executing color matching is activated in the host PC 100, first, in step S310, the LUT created as shown in (1) above is loaded into the memory as color matching initialization processing. . As described above, since the size of this LUT is 64 MB, which is the number of colors to the cube of 256 colors × 4 colors, it is necessary to secure a memory area (static area) for this load. At this time, the image data to be processed is also loaded into the memory.

  In step S320, when the memory area is secured and loading of the LUT and image data is successful, block test processing (details will be described later) in the second embodiment is executed in step S330. By this block test process, the optimum block size is written in the tag portion of the image as described above. Thereafter, in step S340, CMM initialization processing based on the block size is performed to prepare for high-speed color matching processing.

  Hereinafter, the block test process in step S330 will be described in detail. First, in step S331, it is determined whether or not the processing target image data has already been subjected to the block test. If the block test has not been performed, the process proceeds to the block test initialization process in step S332. finish.

  In the block test initialization process in step S332, the type of block to be inspected and the respective block size are determined. Here, since four types of block inspection are performed, each is classified into A, B, C, and D, and the corresponding block size is set. In this case, A = 16 × 16, B = 32 × 32, C = 64 × 64, and D = 128 × 128 are set.

  Then, a test with each block size is sequentially performed. First, as block test A in step S333, color matching processing is performed with the block size set to 16 × 16, and the time taken for processing is measured and recorded. Then, sequentially, block test B is performed in step S334, block test C is performed in step S335, and block test D is performed in step S336. After all the block tests are performed, the test results are totaled in step S337, the block size enabling the fastest processing is determined, and the block size is written in the tag portion of the image.

  Note that the block size inspection process shown in FIG. 6 may be executed as application software that handles an image, or may be executed as an internal process in a digital camera or the like. In any case, the block test result as described above may be preliminarily embedded in the image data as tag information.

  As described above, when image data such as JPEG with the block test result added as tag information is input to the printer controller 200, the rendering unit develops the image. At this time, the CMS process is performed after the JPEG is expanded. The loop method, that is, the size of the processing unit block is switched based on the block size information described above during the CMS process.

  Here, a specific example of speeding up the CMS processing according to the second embodiment will be shown. Compare the processing time of the CMS processing in the conventional color printing system shown in FIG. 12 and the CMS processing in the printing system of the second embodiment shown in FIG. 4 for an image (approximately 48 MB) consisting of 4992 × 3328 pixels. . As a result, according to the second embodiment, the memory development using the LUT of the image data shown in (1) above is first performed to increase the speed by about 8 times (6.56 seconds → 0.825 seconds). Was achieved. Further, by performing the block size control shown in the above (2), a speed increase of about 1.25 times (0.825 seconds → 0.660 seconds) was achieved. That is, by controlling the above (1) and (2) in the CMS processing, the overall processing time was shortened from 6.56 seconds to 0.660 seconds, and about 10 times faster processing was achieved. become.

  As described above, according to the second embodiment, by performing the LUT process (1) and the block size control (2), the CMS process can be dramatically speeded up. Therefore, for example, it is possible to cope with high-speed printing for business use.

  Although the above-described processes (1) and (2) can be executed at a high speed even if they are executed alone, a synergistic effect by executing both processes is greater.

  Of course, the above-described first embodiment can be combined with the second embodiment. That is, when the processing target in the second embodiment is compressed image data, as described in the first embodiment, the cache data to be used is created by multithreading according to the frequency component, thereby further increasing the processing speed. Can be improved.

  In the second embodiment, an example in which CMS processing is performed on the printer controller 200 side has been described. However, it is also possible to execute this on the host PC 100 side, and the processing speed is similarly improved.

<Third Embodiment>
The third embodiment according to the present invention will be described below.

  In the printing system of the second embodiment described above, it is assumed that the data transfer connectivity from the host PC 100 to the printer controller 200 is thin. However, in the third embodiment, it is assumed that the connectivity is sufficiently thick. .

  FIG. 7 is a diagram illustrating a schematic configuration of a printing system according to the third embodiment. FIG. 7 shows a configuration example in which all the color matching processing is performed on the host PC 100 side, and the rendering processing is distributed between the host PC 100 and the printer controller 200, taking advantage of the thick connectivity. A host PC 100 shown in FIG. 7 includes a CPU having a multi-core capable of simultaneously executing a plurality of threads at high speed and a large-capacity main memory (DDR, XDR, etc.). Of course, this CPU also has a cache.

  In FIG. 7, as shown at 410, the following three objects are illustrated as objects in the print job data to be processed.

A command (C1) representing flat fill and its color value (rgb1)
A command (C2) representing a transparent object and its color value (rgb2)
A command (C3) representing an image and its color value (rgb3)
As shown in 420, the job data is subjected to CMS processing for all of C1, C2, and C3 on the host PC 100 side, and C2 and C3 are rendered up to rendering. The job data processed in this way by the host PC 100 is transmitted to the printer controller 200 by wide-band connectivity, where only C1 is rendered as shown at 430.

  Hereinafter, the speeding up of the color matching process in the third embodiment will be described in detail. In the third embodiment, the color matching process is realized by executing a program by the CPU in the host PC 100. In order to improve the execution speed, the same control as in the second embodiment is performed. That is, the CPU in the host PC 100 performs (1) LUT conversion of the interpolation calculation result and (2) block size control. Since these processes have already been described in the second embodiment, they are omitted here. In the third embodiment, the CPU in the host PC 100 further performs the control shown in the following (3).

  (3) Start up multiple threads and parallelize processing.

  Hereinafter, the control (3) will be described in more detail.

  As described above, the host PC 100 constituting this system can execute multi-thread processing, two CPU cores can be used, and four threads can be processed simultaneously. Note that 2-thread processing in one core is pseudo-parallel processing.

  In general, when four threads are executed using two cores, the processing is not necessarily performed at a high speed due to overlapping of internal cache memory and bus utilization time. How many processes can be run to get high-speed processing is not known unless it is actually executed.

  Even in a host PC that appears to be equipped with one CPU, there may actually be two cores in one chip. In this case, when viewed from the software side, it may seem that all four CPUs are mounted, and it is not known how many threads can be executed completely and in parallel by the CPUs.

  Therefore, in the third embodiment, a plurality of threads are actually started up on the host PC 100 side and color matching processing is operated, and a preliminary check is performed as to when the number of threads is highest and processing is completed at the highest speed. It is characterized by that.

  The thread test in the third embodiment will be described in detail using the flowchart of FIG. First, in step S510, a system setting file is loaded. Next, in step S520, it is determined whether or not the thread test has already been completed as a system. If the thread test has not been completed, the process proceeds to the thread test initialization process in step S530, and if completed, the process ends.

  When the thread test initialization process in step S530 is completed, in step S540, a candidate for the number of threads to be tested is selected and the thread test is sequentially started (S551, S552, S553, and S554). Here, it is assumed that four types of threads, 1 thread, 2 threads, 4 threads, and 8 threads, are selected.

  Here, as the target image of the thread test, it is preferable to use an image obtained by performing a block test as described in the second embodiment on a predetermined test image. That is, each thread test process is executed in parallel by the unit of the optimum block size read from the tag information given to the image data.

  In step S560, based on the processing results in steps S551 to S554, the number of threads that realizes the highest speed processing is determined as the optimum number of threads in the host PC 100. In step S570, the optimum thread number is written in the system setting file in the host PC 100, thereby terminating the thread test process. The thread test executed in parallel as steps S551 to S554 is controlled by a thread monitoring control unit (not shown) provided in the host PC 100.

  Hereinafter, the color matching process in the third embodiment will be described in detail with reference to the flowchart of FIG.

  When a program for executing color matching is started in the host PC 100, first, in step S610, an LUT created as shown in (1) in the second embodiment is used as color matching initialization processing. Loaded into memory. As described above, since the size of this LUT is 64 MB, which is the number of colors to the cube of 256 colors × 4 colors, it is necessary to secure a memory area (static area) for this load.

  When the memory area is secured and loading of the LUT is completed, in step S630, the block size test shown in (2) in the second embodiment is executed. Note that the processes (1) and (2) are described as being performed on the printer controller 200 side in the second embodiment, but are assumed to be performed on the host PC 100 side in the third embodiment.

  Thereafter, in step S640, the thread test shown in (3) above is executed. The optimum thread value and block size are determined by the processing of steps S630 and S640.

  In step S650, color matching processing is performed. In this color matching processing, distributed processing according to the number of threads determined in step S640 is executed. Further, when image processing (color matching processing) is performed in each thread, the block size determined in step S630 is adopted as the minimum processing unit.

  As described above, according to the third embodiment, the optimum number of threads for CMS processing is checked in advance in the host PC 100 that enables multi-thread processing. When performing CMS processing on the host PC 100 side, in addition to (1) LUT processing and (2) block size control as in the second embodiment, (3) multi-thread processing is performed. Faster CMS processing is realized.

  Note that it is also effective to execute the color matching processing on the compressed image data shown in the first embodiment described above with the optimum number of threads in the third embodiment.

<Modification>
In each of the above-described embodiments, an example in which an image block that is a CMS processing unit is configured as a square has been shown. However, this is not necessarily optimal, and the block shape is very complicated depending on the image. There is also a possibility. Moreover, you may comprise so that four types of structures, such as 8x8, 16x16, 32x32, 64x64, may be combined.

  For example, as shown in the left part of FIG. 10, a group of images may be configured based on DCT information. According to the figure, it can be seen that the block is divided into two regions. As a result, the data cache of the CPU can be accessed more effectively, and further improvement in processing speed can be expected.

  The right part of FIG. 10 is an example in which blocks are grouped so as to have similar color values when one image is processed by two threads. This figure shows that the processing by two threads is performed for the classification based on the DCT information shown in the left part of FIG. By collecting similar colors and grouping them in this way, the cache hit rate increases, and higher speed processing can be expected.

<Other embodiments>
Although the embodiment has been described in detail above, the present invention can take an embodiment as a system, apparatus, method, program, storage medium (recording medium), or the like. Specifically, the present invention may be applied to a system composed of a plurality of devices (for example, a host computer, an interface device, a printing device, a web application, etc.), or may be applied to a device composed of a single device. good.

  In the present invention, a software program for realizing the functions of the above-described embodiments is supplied directly or remotely to a system or apparatus, and the computer of the system or apparatus reads and executes the supplied program code. Is also achieved. The program in this case is a program corresponding to the flowchart shown in the drawing in the embodiment.

  Accordingly, since the functions of the present invention are implemented by computer, the program code installed in the computer also implements the present invention. In other words, the present invention includes a computer program itself for realizing the functional processing of the present invention.

  In this case, as long as it has a program function, it may be in the form of object code, a program executed by an interpreter, script data supplied to the OS, or the like.

  Recording media for supplying the program include the following media. For example, floppy disk, hard disk, optical disk, magneto-optical disk, MO, CD-ROM, CD-R, CD-RW, magnetic tape, nonvolatile memory card, ROM, DVD (DVD-ROM, DVD- R).

  As a program supply method, the following method is also possible. That is, the browser of the client computer is connected to a home page on the Internet, and the computer program itself (or a compressed file including an automatic installation function) is downloaded from the browser to a recording medium such as a hard disk. It can also be realized by dividing the program code constituting the program of the present invention into a plurality of files and downloading each file from a different homepage. That is, a WWW server that allows a plurality of users to download a program file for realizing the functional processing of the present invention on a computer is also included in the present invention.

  In addition, the program of the present invention is encrypted, stored in a storage medium such as a CD-ROM, distributed to users, and key information for decryption is downloaded from a homepage via the Internet to users who have cleared predetermined conditions. It is also possible to make it. That is, the user can execute the encrypted program by using the key information and install it on the computer.

  Further, the functions of the above-described embodiments are realized by the computer executing the read program. Furthermore, based on the instructions of the program, an OS or the like running on the computer performs part or all of the actual processing, and the functions of the above-described embodiments can also be realized by the processing.

  Further, the program read from the recording medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, and then executed, so that the program of the above-described embodiment can be obtained. Function is realized. That is, based on the instructions of the program, the CPU provided in the function expansion board or function expansion unit can perform part or all of the actual processing.

1 is a diagram illustrating a schematic configuration of an image forming unit according to an embodiment of the present invention. It is a figure which shows schematic structure of the image processing system in this embodiment. It is a figure which shows the image data example processed in this embodiment. It is a figure which shows schematic structure of the printing system in 2nd Embodiment. It is a figure which shows the cache outline incorporated in CPU in the printer controller in 2nd Embodiment. It is a flowchart which shows the block test in 2nd Embodiment. It is a figure which shows the schematic structural example of the color printing system in 3rd Embodiment. It is a flowchart which shows the thread test in 3rd Embodiment. It is a flowchart which shows the color matching process in 3rd Embodiment. It is a figure for demonstrating the modification of this invention. It is a figure which shows the example of schematic structure of the general image processing system using an image compression technique. It is a figure which shows the schematic structural example of a general color printing system.

Claims (18)

A decompression step for decompressing the compressed image data;
A cache generation step of generating cache data for color conversion based on a frequency component for each processing unit block of the image data;
A color conversion step for performing color conversion processing based on the cache data for the decompressed image data;
An image processing method comprising:   The image processing method according to claim 1, wherein, in the cache generation step, the cache data is generated by multithreading.   The image processing method according to claim 1, wherein the cache generation step is executed during the execution of the decompression step.   The image processing method according to claim 1, wherein, in the cache generation step, the cache data is generated based on a DC component for each block.   5. The image processing method according to claim 4, wherein, in the cache generation step, the cache data is generated for a block in which the DC component is larger than a predetermined value.   6. The image processing method according to claim 1, wherein in the cache generation step, the cache data is stored in a global area. An image processing method for performing color conversion processing on image data,
A test step for experimentally executing a color conversion process with a plurality of block sizes on the image data;
Based on the test result in the test step, a block size specifying step for specifying a block size in which the color conversion process is executed in the shortest time;
A size information giving step for giving size information indicating the block size to the image data,
A color conversion process for the image data is performed with a block size indicated by the size information.   8. The image processing method according to claim 7, wherein, in the size information adding step, the size information is added to a tag portion of the image data. An interpolation calculation step for performing an interpolation calculation for color conversion on the target color of the color conversion process;
A result holding step of storing the calculation result in the interpolation calculation step in a file, and
The image processing method according to claim 7, wherein a color conversion process for the image data is performed based on the file.   The image processing method according to claim 9, wherein the file is a lookup table.   11. The image according to claim 10, wherein the look-up table is loaded into an internal memory that enables high-speed access by a CPU that executes the program when a program that executes color conversion processing is started. Processing method. An image processing method for performing color conversion processing on image data in a host computer capable of parallel processing by a plurality of threads,
A test step for experimentally executing the color conversion process with a variable number of threads;
Based on the test result in the test step, the number of threads specifying step for specifying the number of threads that executed the color conversion process in the shortest time;
A thread number information holding step for holding thread number information indicating the number of threads in a system file in the host computer,
A color conversion process for the image data is performed as parallel processing based on the number of threads indicated by the thread number information. Decompression means for decompressing the compressed image data;
Cache generating means for generating cache data for color conversion based on frequency components for each processing unit block of the image data;
Color conversion means for performing color conversion processing based on the cache data for the decompressed image data;
An image processing apparatus comprising: Test means for experimentally executing color conversion processing with a plurality of block sizes on image data;
Based on the test result by the test means, a block size specifying means for specifying the block size that has executed the color conversion processing in the shortest time;
Size information giving means for giving size information indicating the block size to the image data;
Color conversion means for performing color conversion processing on the image data according to the block size indicated by the size information;
An image processing apparatus comprising: further,
Interpolation calculation means for executing interpolation calculation for color conversion for all colors to be converted;
And a result holding means for storing the calculation result by the interpolation calculation means in a file,
The image processing apparatus according to claim 14, wherein the color conversion unit performs a color conversion process based on the file on the image data. An image processing system for performing color conversion processing on image data in a host computer capable of parallel processing by a plurality of threads,
Test means for performing the color conversion process on a trial basis with a variable number of threads;
As a test result in the test step, thread number specifying means for specifying the number of threads capable of executing color conversion processing in the shortest time;
Thread number information holding means for holding thread number information indicating the number of threads in a system file in the host computer,
The image processing system according to any one of claims 1 to 12, wherein the color conversion processing for the image data is performed as parallel processing based on the number of threads indicated by the thread number information.   A program that, when executed on a computer, causes the computer to execute the image processing method according to any one of claims 1 to 12.   A recording medium on which the program according to claim 17 is recorded.

Images