JP3582540B2 - Resolution conversion device and resolution conversion method - Google Patents

Description

[0001]
[Industrial applications]
The present invention relates to a resolution conversion apparatus that converts digital image information of a still image or a moving image composed of characters, graphics, and natural images into image information having different numbers of pixels and displays the image information on a display. The present invention relates to a resolution conversion device and a resolution conversion method suitable for displaying input image information in a display device having a different resolution or a window environment in moving image display.
[0002]
[Prior art]
When converting the resolution of a digital image to increase the number of pixels of the original image, the original image is enlarged and displayed on the same display device. Therefore, it is effective for enlarged display of a window. In addition, it is possible to display on a display device having a higher resolution than the resolution of the original image. Therefore, the simplest method is to increase the number of pixels in the original image by filling the same pixel in the horizontal and vertical directions for one pixel of the original image.
[0003]
Japanese Patent Application Laid-Open No. 6-124189 discloses an "image display device and an image display control method" in which pixel data in the horizontal direction and the vertical direction of an original image are averaged, and the number of pixels of the original image is reduced. There are ways to reduce this. In this method, a line is drawn by using a line drawing algorithm known as a conventional technique, and a straight line of an original image is represented by y horizontal components of at least one dot. Degrees or less. The length of each horizontal component can be considered to represent the number of pixel data of image input corresponding to one pixel data after reduced scaling written in the VRAM. Although the length of each horizontal component may be different or the same in all cases, reduction scaling in the horizontal direction is performed by averaging pixel data for n dots. The same applies to the vertical direction. By making the inclination of this straight line 45 degrees or more, enlargement processing becomes possible, and enlargement scaling in the horizontal and vertical directions is performed.
[0004]
Further, instead of simply filling the same pixel, when filling the same pixel, as shown in FIGS. 18 and 22, the gradation values of the interpolation pixels are connected so that the gradation values between the pixels of the original image are connected by a straight line. Conventionally, there is a method of increasing the number of pixels of the original image by calculating and filling the pixels. FIGS. 18 and 22 show the gradation values at each pixel position in the horizontal direction.
[0005]
FIG. 17 shows a method for arbitrarily converting the resolution of an original image. Reference numeral 1701 denotes original image data, reference numeral 1702 denotes processing for vertically multiplying the original image by L1 times and width to M1 times, and reference numeral 1703 denotes enlarged image data. Reference numeral 1704 denotes processing for expanding the enlarged image data by 1 / L2 times vertically and 1 / M2 times horizontally, and 1705 denotes reduced image data. Here, L1, L2, M1 and M2 are positive integers. This processing corresponds to image enlargement if L1> L2, M1> M2, and image reduction if L1 <L2, M1 <M2. As described above, the image size can be more freely converted by combining the image size enlargement process and the image size reduction process.
[0006]
[Problems to be solved by the invention]
<< Problem 1 >> In the case of increasing the number of pixels of the original image according to the conventional method, the method of increasing the number of pixels of the original image by filling the same pixel in the horizontal direction and the vertical direction with respect to one pixel of the original image includes: A mosaic-like jagged image becomes conspicuous, and there is a problem that image quality is deteriorated.
[0007]
<< Problem 2 >> When increasing the number of pixels of the original image according to the above-described conventional method, the same pixels are not simply filled, but when the same pixels are filled, as shown in FIGS. In the method of calculating the gradation value of the interpolation pixel so as to connect the tonal values with a straight line and increasing the number of pixels of the original image by filling the pixels, the image does not become a mosaic-like jagged image, but the boundary of shading There is a problem that high-frequency components such as contour portions of an image in which the image is clear are smoothed, and the entire image is blurred.
[0008]
<< Problem 3 >> In the case of increasing the number of pixels of the original image according to the above-described conventional method, in a method of more freely converting the image size by combining the image size enlargement process and the image size reduction process, the image enlargement process is performed. , The resolution conversion process takes a long time.
[0009]
Therefore, a first object of the present invention is to prevent a mosaic jagged image from being noticeable in a resolution conversion for increasing the number of pixels of an original image and realize a high-quality image.
[0010]
Further, a second object of the present invention is to realize high-quality images by preserving high-frequency components such as outline portions of images having sharp gray-scale boundaries in resolution conversion for increasing the number of pixels of an original image. is there.
[0011]
Further, a third object of the present invention is to realize high-speed resolution conversion processing that satisfies digital moving image processing in resolution conversion for increasing the number of pixels of an original image.
[0012]
[Means for Solving the Problems]
In order to achieve the object of the present invention, in a resolution conversion device that converts a digital original image into a digital image having a different number of pixels according to a specified conversion magnification,
For each block of a predetermined number of pixels of the digital original image, determining means for determining an interpolation position of an interpolation pixel to be interpolated with respect to the block of the digital original image according to the conversion magnification,
For each block of the digital original image, according to a predetermined interpolation formula in which a coefficient is determined by a pixel position and pixel data of each pixel of the block, a pixel to be interpolated to the interpolation position determined by the determination unit is determined. A generating unit that generates image data and outputs a converted digital image from the image data of the pixels of the digital original image and the generated image data of the pixels to be interpolated.
[0013]
Further, in a resolution conversion device that converts a digital original image into a digital image having a different number of pixels according to a designated conversion magnification,
For each block of a predetermined number of pixels of the digital original image, determining means for determining an interpolation position of an interpolation pixel to be interpolated with respect to the block of the digital original image according to the conversion magnification,
For each block of the digital original image, the pixel position and pixel data of each pixel of the block, and according to a predetermined interpolation formula whose coefficient is determined by the pixel position and pixel data of the pixel adjacent to the block, Generating image data of a pixel to be interpolated at the interpolation position determined by the determining unit, and outputting a digital image after conversion from the image data of the pixel of the digital original image and the generated image data of the pixel to be interpolated; And a generation unit.
[0014]
[Action]
In the resolution conversion apparatus according to the present invention, the determination unit determines, for each block of a predetermined number of pixels of the digital original image, an interpolation position of an interpolation pixel to be interpolated with respect to the block of the digital original image according to the conversion magnification. I do. In this case, for example, an interpolation position can be defined in a table for each conversion magnification, and the interpolation position can be determined according to the conversion magnification with reference to this table. Also, the unit block can be a block of 8 × 8 pixels.
[0015]
The generating unit interpolates, for each block of the digital original image, the interpolation position determined by the determination unit according to a predetermined interpolation formula in which a coefficient is determined by a pixel position and pixel data of each pixel of the block. The image data of the pixel to be converted is generated, and the converted digital image is output from the image data of the pixel of the digital original image and the generated image data of the pixel to be interpolated. The interpolation formula can be a spline function or a Bezier function.
[0016]
Thereby, the pixel data of the interpolation pixel can be optimized.
[0017]
According to the present invention, in the resolution conversion for increasing the number of pixels of the original image, it is possible to prevent a mosaic jagged image from being conspicuous, and realize a high-quality image with respect to a natural image or the like whose contour is not clear. In addition, for characters and figures with clear contours, high-frequency components such as contours of the image with sharp boundaries are preserved, and high-quality images can be realized.
[0018]
When an interpolation position is provided between pixels of a block to be interpolated and an adjacent block, in addition to the pixel position and pixel data of the block, the pixel position and pixel data of a pixel adjacent to the block to be interpolated are further determined. Can be used to determine the coefficients of the interpolation formula. Accordingly, an interpolated pixel can be generated between pixels in a portion where the unit block is adjacent to the unit block, that is, in a position outside the pixel of the unit block, and pixel data is also discontinuous in the adjacent portion of the unit block. Can be avoided.
[0019]
Further, the determining unit sets the conversion magnification as a conversion ratio between a horizontal direction and a vertical direction, and the generation unit generates the image data of the pixel to be interpolated in the horizontal direction. In the case where the apparatus is provided with the vertical direction generating means for performing the direction, the generation can be performed in each of the generating means, so that the speed can be increased.
[0020]
【Example】
Hereinafter, embodiments of the present invention will be described in detail.
[0021]
First, a general configuration and a processing flow of the present invention will be described, and then, a detailed configuration and a processing flow for solving the problem will be described.
[0022]
In the present embodiment, an interpolated pixel is generated, the number of pixels is increased, and the image is converted into an enlarged image.For each unit pixel block having a predetermined number of pixels, the number of pixels to be interpolated according to the magnification and the number of pixels are interpolated. The image data of the pixel to be interpolated at the interpolation position is generated according to a predetermined interpolation formula in which a coefficient is determined based on the pixel position of the block and the pixel data. That is, when the gradation value of the pixel in the unit block is represented by a predetermined function (interpolation formula), the gradation value of the interpolated pixel is obtained by the gradation value on the function for the position of the pixel to be interpolated.
[0023]
FIG. 1 shows a block diagram of a hardware configuration for converting a digital original image into a digital image having a different number of pixels. FIG. 2 shows how pixel conversion is performed. The device in this embodiment need not be limited to the configuration as shown in FIG. 1 and may be a single element formed as an LSI as a microcomputer peripheral device or a device systemized by software. Here, the digital original image data such as a still image or a moving image is obtained by expanding compressed data such as JPEG or MPEG or digitizing a signal such as NTSC.
[0024]
In FIG. 1, a horizontal scaling operation unit 101 converts the number of pixels in the input original image data in the horizontal direction. The horizontal direction scaling operation unit 101 inputs the original image data 103 of eight pixels for one line in the horizontal direction with respect to the image data obtained by dividing the original image data into unit blocks of 8 × 8 pixels, and Perform the conversion in order. Further, the horizontal scaling operation unit 101 inputs a control signal 105 indicating a horizontal scaling factor for indicating how many pixels are added to the original image in the horizontal direction. As shown in FIG. 2, the horizontal scaling operation unit 101 inputs the original image data 202 for eight pixels in one line in the horizontal direction among the original image data divided into unit blocks 201 of 8 × 8 pixels, The image data 203 for 8 ± n pixels per line is output. In FIG. 2, hatched pixels C <b> 1 and C <b> 2 in the image data 203 of 8 ± n pixels per line are interpolation pixels generated by the horizontal scaling operation unit 101.
[0025]
Next, the vertical scaling operation unit 102 converts the number of pixels in the vertical direction with respect to the image data output from the horizontal scaling operation unit 101 and converted in the horizontal direction. The vertical scaling operation unit 102 inputs the image data 104 of 8 + n pixels for one line in the horizontal direction outputted by the horizontal scaling operation unit 101 for eight lines one by one, and the image data 107 for 8 + n lines in the vertical direction. Is output. In addition, the vertical direction scaling operation unit 102 inputs a control signal 106 indicating a vertical scale factor for indicating how many pixels are added to the original image in the vertical direction. As shown in FIG. 2, the vertical scaling operation unit 102 inputs eight lines of image data 203 for one line 8 + n pixels whose number of pixels has been converted in the horizontal direction, and outputs image data 204 whose number of pixels has been converted in the vertical direction. Is output. In FIG. 2, hatched pixels P1 to P10 are interpolated pixel data generated by the vertical scaling operation unit 102. The vertical scaling operation unit 102 generates interpolation lines 205 and 206, and outputs 8 × 8 pixels. The number of pixels is converted from the original image data divided into the unit blocks of the above into image data 207 of the unit block of 10 × 10 pixels.
[0026]
FIG. 7 shows a system configuration example to which the present invention is applied. The CPU 701 is a central processing unit that processes information. An image data input unit 702 inputs compressed digital still image data or moving image data. The buffer memory 703 stores image data. For example, when the image data is MPEG-compressed moving image data, the buffer memory 703 is used to store image data before and after decompression when decompressing. Has a capacity to store at least 8 × 8 pixels. The image processing LSI 704 performs resolution conversion as shown in FIG. The VRAM 705 stores the image data after the resolution conversion. The display control unit 706 reads out the image data after the resolution conversion stored in the VRAM 705 and displays the image data on the liquid crystal panel. D / A 708 is a digital / analog converter for audio output. The system bus 709 includes a data bus, an address bus, and a control bus. The input unit 710 receives an instruction such as a resolution conversion magnification.
[0027]
In the present system configuration, the CPU 701, the buffer memory 703, the image processing LSI 704 for performing the scaling of the present invention, and the display control 706 may be a single element made into one LSI.
[0028]
Next, the details of the resolution conversion will be described with reference to FIGS. FIG. 3 shows a processing flow in the resolution conversion of the present embodiment.
[0029]
In FIG. 3, the magnification of the resolution conversion for determining how many times the original image is to be multiplied in the horizontal and vertical directions is received by the input unit 710 of the apparatus shown in FIG. 7 (S301). The CPU 701 sets the resolution conversion magnification data in a parameter register for setting the parameters of the image processing LSI 704. The image data input unit 702 receives the image data, divides it into 8 × 8 pixel unit blocks, and stores the image data in the buffer memory 703 for each of the divided unit blocks. Next, the image processing LSI 704 inputs 8-pixel data of one line in the horizontal direction of original image data divided into unit blocks of 8 × 8 pixels stored in the buffer memory 703 (S302). The image processing LSI 704 generates an interpolated pixel in the horizontal direction with respect to the input eight-pixel data for one horizontal line, and performs a scaling operation for eight horizontal lines (S303). Next, the image processing LSI 704 generates interpolation pixels in the vertical direction with respect to the eight lines of data that have been subjected to the horizontal scaling operation, generates an interpolation line at a time, and performs the vertical scaling operation. (S304). Finally, the data is written to the VRAM 705, which is a display memory, for each unit pixel block after the horizontal and vertical scaling operations (S305).
[0030]
Next, the number of pixels to be interpolated according to the magnification and the position at which the pixels are interpolated will be described in detail.
[0031]
FIG. 4 is a diagram illustrating a relationship between a horizontal scaling factor for the horizontal scaling operation unit 101 illustrated in FIG. 1 to generate an interpolation pixel in the horizontal direction and perform a horizontal scaling operation process, and a corresponding number of interpolation pixel generation points. 5 shows an example of a horizontal direction interpolation pixel generation table showing the following. Since an interpolation pixel is generated for each horizontal line of the original image data divided into unit blocks of 8 × 8 pixels with respect to the original image data, the number of input pixels is eight. For example, in the case of 8 bits for each of RGB in color display, the data length is 24 bits per pixel. The image data of each pixel may be YUV digital data represented by luminance and color difference. As shown in FIG. 4, scaling in the horizontal direction is realized by increasing the number of generated pixel generation points by one for each magnification for eight input pixels. For example, when the horizontal magnification is set to 1.625, the number of generated interpolation pixels is 5, and the number of output pixels is 13. The number of interpolation pixels to be generated can be specified from the horizontal magnification, but the position at which the interpolation pixels are generated affects the image quality. The position where the interpolation pixel is generated will be described later.
[0032]
FIG. 5 shows the vertical scaling factor for the vertical scaling operation unit 102 shown in FIG. 1 to generate interpolation pixels in the vertical direction and perform vertical scaling operation processing, and the number of interpolation pixel generation lines corresponding thereto. 6 shows an example of a vertical direction interpolation pixel generation table showing the relationship. Since the interpolation pixels are generated in the vertical direction of the original image data divided into unit blocks of 8 × 8 pixels with respect to the original image data and the interpolation lines are generated, the number of input lines is eight. As shown in FIG. 5, the scaling in the vertical direction is realized by increasing the number of generated lines by one for each magnification for eight input lines. For example, when the magnification in the vertical direction is set to 1.375 times, the number of generated lines of the interpolation pixel is three, and the number of output lines is eleven. As in the case of horizontal scaling, the number of lines for generating interpolation pixels can be specified based on the magnification in the vertical direction.
[0033]
Therefore, referring to FIG. 6, the number of interpolation pixel generation points is specified based on the horizontal magnification, and the generation position of the interpolation pixel will be described.
[0034]
FIG. 6 shows a case where interpolation pixels are generated from original image data of eight pixels per line in the horizontal direction divided into unit blocks of 8 × 8 pixels with respect to the original image data. It is illustrated for each case. The generation position of the interpolation pixel is not limited to this, and the user may be allowed to specify the generation position by himself. In FIG. 6, a rectangular portion of a white frame indicates each pixel of 8 pixels per line of the original image data divided into unit blocks of 8 × 8 pixels. The hatched quadrangular portion is an interpolated pixel, and indicates the generation position. n is an integer equal to or greater than 0, and indicates the number of occurrence points of the interpolation pixel. Reference numeral 600 denotes a state in which the number of occurrence points of the interpolation pixels is n = 0 and pixels 1 and 2 of the original image are arranged in order in the horizontal direction up to pixel 8. Reference numeral 601 indicates that the number of generated interpolation pixels is n = 1, and one interpolation pixel is generated between pixel 1 and pixel 2 of the original image in the horizontal direction. Numeral 602 indicates that the number of interpolated pixels is n = 2, one interpolated pixel is generated between pixel 1 and pixel 2 of the original image in the horizontal direction, and the interpolated pixel is between pixel 5 and pixel 6. For one pixel. The horizontal magnification at this time is 1.25, as shown in the horizontal interpolation pixel generation table in FIG. Further, when the horizontal magnification is increased, in 608, the number of generated interpolation pixels is n = 8, and two interpolation pixels are generated between pixel 1 and pixel 2 of the original image in the horizontal direction. One interpolated pixel is generated between pixel 2 and pixel 3, one interpolated pixel is generated between pixel 3 and pixel 4, and one interpolated pixel is generated between pixel 4 and pixel 5. , An interpolated pixel is generated between pixel 5 and pixel 6 for one pixel, an interpolated pixel is generated for one pixel between pixel 6 and pixel 7, and an interpolated pixel is generated between pixel 7 and pixel 8. One pixel is generated. The horizontal magnification at this time is twice as shown in the horizontal interpolation pixel generation table in FIG. Similarly, an interpolated pixel is generated with respect to the number n of interpolated pixel occurrence points as indicated by 603 to 624.
[0035]
As described above, the number of interpolation pixels to be generated and the generation position can be defined in advance according to the magnification.
[0036]
Further, the above description relates to the generation position of the interpolation pixel in the horizontal direction. However, the interpolation pixel in the vertical direction may be similarly replaced with the interpolation line. That is, the number of generation lines of the interpolation pixel can be defined based on the magnification in the vertical direction, and the generation position of the line of the interpolation pixel can be similarly defined.
[0037]
Next, the horizontal and vertical scaling operation processing method will be specifically described. FIG. 8 shows a flowchart in the case where the horizontal scaling operation unit 101 shown in FIG. 1 enlarges the image by 1.25 times in the horizontal direction. In the case of multiplying by 1.25, as shown in FIG. 4, the number of generated points is n = 2, and the pixel data of eight pixels per line is converted into ten pixels.
[0038]
In FIG. 8, first, the horizontal scaling operation unit 101 secures a memory (array) area for storing pixel data of 10 pixels, which is the number of pixels after pixel number conversion (S801), and data of one line Is input and stored in the secured memory (S802). At this time, the interpolation pixels are stored according to the horizontal interpolation pixel generation tables shown in FIGS. Then, a curve based on a spline function is generated between the pixel data of eight pixels input in S803, an interpolation pixel is generated, and the interpolation pixel is stored in the memory area (array) that has been previously opened.
[0039]
A method of generating this curve will be described with reference to FIG. FIG. 9 shows pixel data for a pixel position when the number of pixels is converted by horizontal scaling operation processing. The horizontal axis represents the position (order) of the pixel data, and the vertical axis represents the value (luminance) of the pixel data.
[0040]
In FIG. 9, reference numeral 901 denotes pixel data of eight pixels before pixel number conversion, and reference numeral 902 denotes pixel data of ten pixels after pixel number conversion. In S802 shown in FIG. 8 described above, pixel data for eight pixels is arranged with a position where an interpolation pixel is inserted, as shown at 903 in FIG. 9, according to the horizontal interpolation pixel generation tables shown in FIGS. Then, as shown by 904 in FIG. 9, a curve 905 is generated from the eight pixel data by a spline function. Here, the spline function is a piecewise polynomial function as shown in FIG. 20, and is defined by different polynomial curves of at most m degrees within a small section, and these are connected as smoothly as possible to each other. Things. The values at the intersections between the curve 905 and the positions C1 and C2 where the interpolation pixels are generated are the values (luminance) 906 and 907 of the interpolation pixel data. The horizontal scaling operation unit 101 generates the interpolation pixel data 906 and 907 in this manner, ends the scaling operation processing for one line, and outputs the pixel data 902 for ten pixels per line to the vertical scaling operation unit 102. I do. As described above, the horizontal scaling operation processing unit 101 sequentially performs the scaling operation processing on a line-by-line basis, and outputs the result to the vertical direction scaling operation processing unit 102.
[0041]
The vertical resolution conversion processing will be described with reference to FIGS. FIG. 10 shows a flowchart in the case where the vertical scaling operation unit 102 shown in FIG. 1 enlarges the image vertically by 1.25 times. FIG. 11 shows pixel data for a pixel position when the number of pixels is converted by vertical scaling operation processing. In the case of multiplying by 1.25, as shown in FIG. 4, the number of generated lines is n = 2, and the pixel data of eight lines is converted into the pixel data of ten lines.
[0042]
A memory for (number of pixels 8 after enlargement in the horizontal direction + n) × (8 lines) is secured in advance in the vertical scaling operation processing unit 102. Here, since the number of generated points is n = 2, 10 × A memory for 8 = 80 pixels is secured. As shown in FIG. 11, the vertical scaling operation processing unit 102 stores, in the memory, pixel data 1101 to 1108 for eight lines after the horizontal pixel number conversion, which are sequentially sent from the horizontal direction scaling operation unit 101. The data is stored, and when eight lines are stored, the scaling operation in the vertical direction is started.
[0043]
In FIG. 10, the vertical scaling operation processing unit 102 stores pixel data for each column of eight lines stored in the memory, for example, pixel data 1112 of eight pixels for each column as shown in FIG. An input is made (S1001). Then, the vertical scaling operation processing unit 102 generates interpolation pixel data for each column (S1002). In FIG. 11, X ₁₁ , X ₂₁ , X ₃₁ , X ₄₁ , X ₅₁ , X ₆₁ , X ₇₁ , X ₈₁ , A curve 1109 based on a spline function is obtained from eight pixels of pixel data 1112 corresponding to the same column in the vertical direction. At this time, a place where an interpolation line is inserted according to the vertical direction interpolation pixel generation table shown in FIG. 5, that is, between one line and two lines (D ₁₁ ) And between lines 5 and 6 (D ₂₁ ) And generate an interpolation pixel. At this time, the curve 1109 and D ₁₁ And D ₂₁ At the intersections 1110 and 1111 are the values of the interpolated pixel data. These X ₁₁ , X ₂₁ , X ₃₁ , X ₄₁ , X ₅₁ , X ₆₁ , X ₇₁ , X ₈₁ And interpolation pixel data D ₁₁ , D ₂₁ Is written directly into the VRAM, and the next C ₁₁ ~ C ₈₁ And generates the interpolation pixel data in the same manner. FIG. 12 shows the pixel data of the unit block after the pixel number conversion. As shown in FIG. ₁₁ ~ C ₈₁ 1201, C ₁₂ ~ C ₈₂ 1202, D ₁₁ ~ D ₂₀ 1203, D ₂₁ ~ D ₃₀ Interpolated pixel data of 1204 is generated. In this way, interpolated pixel data for 10 pixels in the horizontal direction is generated for each column, 8 × 8 pixels are converted into a unit block of 10 × 10 pixels, and the scaling operation process ends. This process is repeated for the resolution of the original image data. For example, if the resolution of the original image data is 160 × 120 pixel data, there are 20 × 15 = 300 blocks in 8 × 8 pixel units as shown in FIG. The arithmetic processing may be repeated 300 times. Thereby, the image data is converted into image data of 200 × 150 pixels.
[0044]
Next, a specific example will be described in detail. FIG. 23A shows original image data of “A” of 8 × 8 pixels. The image data is 8-bit data (0 to 255), and the numerical values in the figure represent the luminance values of the pixels. When this image data is enlarged by a factor of 1.25, an enlarged image in which interpolation pixels are generated by the conventional linear interpolation method is shown in FIG. 23 (2), and interpolation pixels are generated by curve interpolation using the spline function of the present embodiment. The enlarged image obtained is shown in FIG. Comparing the luminance values of the interpolated pixels after the enlargement, the interpolated pixels generated by the curve interpolation method using the spline function of (3) are more likely to have luminance values of adjacent pixels than the interpolated pixels generated by the linear interpolation method of (2). It turns out that it is close. Therefore, when the interpolation pixels are generated by the curve interpolation method using the spline function, it is possible to realize an image with little blurring even after the enlargement processing in the contour portion of the image.
[0045]
According to this embodiment, the number of pixels to be interpolated and the interpolation position can be defined in advance according to the magnification. Further, a curve based on a spline function is obtained from the image data of the pixels in the unit block in the horizontal direction or the vertical direction, and the image data of the interpolation pixel can be determined on the obtained curve. According to this embodiment, in resolution conversion for increasing the number of pixels of an original image, an image in which mosaic jaggies are conspicuous can be prevented, and a high-quality image can be realized. Further, high-speed resolution conversion processing that satisfies digital moving image processing can be realized. Also, an arbitrary resolution conversion process can be realized with a minimum hardware scale.
[0046]
Next, a second embodiment will be described. In the first embodiment, the case where the interpolation pixel is generated within the pixel of the unit block has been described. However, in the second embodiment, the unit block and the unit block are located between adjacent pixels, that is, the unit block. Interpolated pixels are generated at positions outside the pixels of the block so that the pixel data does not become discontinuous even in an adjacent portion of the unit block.
[0047]
FIG. 14 shows an example of the pattern of the generation positions of the interpolated pixels in the second embodiment when the 8-pixel data is enlarged three times in the horizontal direction. FIG. 14 shows a case where interpolated pixel data 1401 to 1408 are generated from image data 1 to 8 of an original image based on a cubic spline curve 1409. In this case, the interpolation pixel data 1408 is generated outside the original image data 8. When generating interpolated pixel data by curve interpolation using a spline function, to generate a cubic spline curve 1409 from the original pixel data 1 to 8, the interpolation pixel data 1401 to 1407 sandwiched between the original pixel data 1 to 8 are generated. Can be generated, but since the interpolated pixel data 1408 is not sandwiched between the original pixel data 1 to 8, the curve interpolation cannot be performed unless the next point 1410 exists. If the value of the pixel data at the next point 1410 is used as the value of the pixel data of the pixel 7 in order to generate the interpolation pixel by the curve interpolation method, the pixel data of the next input 8 × 8 pixel is Joints may be discontinuous.
[0048]
Accordingly, in the present embodiment, as shown in FIG. 15, when there are pixel blocks 1 to 300, when generating horizontal and vertical interpolated pixel data, a pixel block 1 in units of 8 × 8 pixels, A curve based on a spline function is obtained from the pixel data 1501 of the leftmost column of the pixel block 2 on the right side, the pixel data 1502 of the line on the upper end of the pixel block 21, and the pixel data 1503 on the upper left of the pixel block 22. The scaling processing of the pixel block 1 is performed by interpolation.
[0049]
This processing will be described with reference to the flowchart shown in FIG. The flowchart illustrated in FIG. 16 illustrates a processing procedure in a unit block of a scaling process when the original pixel data of 160 × 120 pixels (300 unit blocks) is enlarged by m × n times.
[0050]
In FIG. 16, the horizontal scaling operation processing unit 101 illustrated in FIG. 1 includes the block 1 illustrated in FIG. 15, the pixel data 1501 of the leftmost one column of the pixel block 2 on the right side thereof, and the pixel data 21 of the pixel block 21. A memory (memory 1) for storing at least the pixel data 1502 of the top line and the pixel data 1503 of the top left of the pixel block 22 is secured. When pixel data is sequentially output for each block, a memory (memory) for storing original pixel data of pixels (8 × 8 × 22 = 1408 pixels) of 22 blocks from block 1 to block 22 1) may be secured (S1601) and data may be stored. In addition, the vertical scaling operation processing unit 102 serves as a memory for storing the result of the vertical scaling operation, the pixels of the blocks 1 to 20 after the horizontal conversion and the pixels of the first line of the blocks 21 to 40 ｛8 × ( A memory (memory 2) for storing original pixel data for (horizontal magnification m)｝ × (8 + 1) is secured (S1602).
[0051]
Then, the horizontal scaling operation processing unit 101 reads the pixel data of the block 1 from the memory 1 and the pixel data of the leftmost column of the block 2, and performs the horizontal scaling operation processing on the original pixel data of 9 pixels per line. Is performed for each line, and the horizontal scaling operation result thus obtained is stored in the memory 2 (S1603). Further, the horizontal scaling operation processing unit 101 reads pixel data of a total of 9 pixels, that is, pixel data 1502 of the upper end one row of the block 21 and pixel data 1503 of the upper left end of the block 22 from the memory 1, and reads the pixel data of the 9 pixels. The horizontal scaling operation is performed from the pixel data of (1) and stored in the memory 2 (S1604). As a result, as shown in FIG. 19, nine lines of data that have been horizontally enlarged by a factor of m are stored in the memory 2. At this time, the pixel data 1501 for the leftmost column of the block 2 corresponding to the ninth pixel of the original pixel data is not stored in the memory 2. Then, the vertical scaling operation processing unit 102 performs vertical scaling operation processing on each column in the same manner as in the horizontal direction from the horizontal operation result stored in the memory 2 and writes the result in the VRAM shown in FIG. (S1605). Thus, the enlargement processing of block 1 is completed. Next, the enlargement processing of block 2 is performed in the same manner. In this way, the enlarging process is performed for each block, and the enlarging process is performed for all blocks.
[0052]
In addition, in the case of a block at the right end, such as 20 block or 40 block, there is no 21st block on the right side in 1504 shown in FIG. 15, so pixel data 1501 for one column at the left end shown in FIG. Pixel data corresponding to the pixel data 1503 at the upper left corner of the pixel block 22 may not exist. In this case, for example, the horizontal scaling processing by curve interpolation is performed by moving the value of the image data corresponding to the seventh pixel of the block 20 in parallel to obtain the value of the image data of the ninth pixel. Is also good. Similarly, in the vertical direction, in block 1504 shown in FIG. 15, there is no block below block 281 to block 300, and thus there is no pixel data corresponding to pixel data 1502 in the top line shown in FIG. 15. Therefore, by performing parallel translation of the value of the image data corresponding to the seventh line of the blocks 281 to 300 to obtain the value of the image data of the ninth line, vertical scaling processing by curve interpolation can be performed.
[0053]
According to the present embodiment, an interpolated pixel can be generated between pixels in a portion where a unit block and a unit block are adjacent to each other, that is, at a position outside the pixel of the unit block. Data can be prevented from being discontinuous.
[0054]
In the above-described first and second embodiments, when performing the curve interpolation, the curve interpolation using the spline function as shown in FIG. 20 was performed. However, the spline function is not limited to this, and the mixed order obtained by simplifying the m-order spline function is used. A spline function or a base spline function may be used, or, as shown in FIG. 21, curve interpolation may be performed using a cubic Bezier curve passing only from the four interpolation points to the start and end points. Interpolation may be performed using another interpolation formula.
[0055]
【The invention's effect】
According to the present invention, a high-quality image can be reproduced by a resolution conversion apparatus and method for converting a digital original image into a digital image having a different number of pixels and displaying the digital image on a display. Further, the resolution conversion processing can be speeded up and reduced in cost.
[Brief description of the drawings]
FIG. 1 is a block diagram of an image enlargement scaling processing method.
FIG. 2 is an explanatory diagram of an image enlargement scaling processing method.
FIG. 3 is a flowchart of an image enlargement scaling processing method.
FIG. 4 is an explanatory diagram of a horizontal interpolation pixel generation table.
FIG. 5 is an explanatory diagram of a vertical direction interpolation pixel generation table.
FIG. 6 is an explanatory diagram of an image enlargement scaling processing method.
FIG. 7 is a system configuration diagram in an image enlargement scaling process.
FIG. 8 is a flowchart of a horizontal scaling operation process.
FIG. 9 is a diagram illustrating generation of interpolated pixel data in horizontal scaling operation processing.
FIG. 10 is a flowchart of a vertical scaling operation process.
FIG. 11 is an explanatory diagram of generation of interpolation pixel data in vertical scaling operation processing.
FIG. 12 is a schematic diagram of pixel data after the number of pixels has been converted to 1.25 times.
FIG. 13 is an explanatory diagram of block division of 8 × 8 pixels in pixel data of 160 × 120 pixels.
FIG. 14 is an explanatory diagram of an image enlargement scaling processing method.
FIG. 15 is an explanatory diagram of an image enlargement scaling processing method.
FIG. 16 is a flowchart illustrating an image enlargement scaling processing method.
FIG. 17 is a block diagram of a conventional image enlargement scaling processing method.
FIG. 18 is a diagram illustrating generation of interpolated pixel data by linear interpolation.
FIG. 19 is a diagram illustrating the structure of a memory 2.
FIG. 20 is a schematic diagram of an m-order spline function.
FIG. 21 is a schematic diagram of a cubic Bezier function.
FIG. 22 is a schematic diagram of linear interpolation.
FIG. 23 is an example of enlarged image data.
[Explanation of symbols]
Reference numeral 101: horizontal scaling calculator, 102: vertical scaling calculator, 701: CPU, 702: image input unit, 703: buffer memory, 704: scaling controller, 705: VRAM, 706: display controller, 707: liquid crystal panel.

Claims (5)

In a resolution conversion device that converts a digital original image into a digital image having a different number of pixels according to a specified conversion magnification,
For each block of a predetermined number of pixels of the digital original image, determining means for determining an interpolation position of an interpolation pixel to be interpolated with respect to the block of the digital original image according to the conversion magnification,
For each block of the digital original image, according to a predetermined interpolation formula in which a coefficient is determined by a pixel position and pixel data of each pixel of the block, a pixel to be interpolated to the interpolation position determined by the determination unit is determined. A generating unit that generates image data and outputs a digital image after conversion from the image data of the pixel of the digital original image and the image data of the generated pixel to be interpolated ,
The determining means determines that at least one of the interpolation positions is provided between pixels of a block to be interpolated and an adjacent block,
The generation means generates the pixel data of the pixel to be interpolated by determining a coefficient of the interpolation formula further using a pixel position and pixel data of a pixel adjacent to the block to be interpolated. Resolution converter. 2. The resolution conversion device according to claim 1, wherein the interpolation formula is a spline function. 2. The resolution conversion device according to claim 1, wherein the interpolation formula is a Bezier function. 2. The method according to claim 1, wherein the determining unit sets the conversion magnification as a conversion magnification between a horizontal direction and a vertical direction.
The resolution conversion apparatus according to claim 1, wherein the generation unit includes a horizontal direction generation unit that generates the image data of the pixel to be interpolated in the horizontal direction and a vertical direction generation unit that performs the vertical direction. In a resolution conversion device that converts a digital original image into a digital image having a different number of pixels according to a specified conversion magnification,
For each block of a predetermined number of pixels of the digital original image, determining means for determining an interpolation position of an interpolation pixel to be interpolated with respect to the block of the digital original image according to the conversion magnification,
For each block of the digital original image, the pixel position and pixel data of each pixel of the block, and according to a predetermined interpolation formula whose coefficient is determined by the pixel position and pixel data of the pixel adjacent to the block, Generating image data of a pixel to be interpolated at the interpolation position determined by the determining unit, and outputting a digital image after conversion from the image data of the pixel of the digital original image and the generated image data of the pixel to be interpolated; A resolution conversion device comprising: a generation unit.

Images