JP3692942B2 - Image processing apparatus, image display apparatus, and image processing method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
  The present invention relates to an image processing apparatus, an image display apparatus, an image processing method, and an image display method for enlarging or reducing a digital image at an arbitrary magnification.
[0002]
[Prior art]
  FIG. 14 is a diagram illustrating a case where the number of pixels is converted to three times by a conventional pixel number conversion method. In the figure, the horizontal axis indicates the horizontal position or vertical position of the image, and the vertical axis indicates the level (brightness) of the image data. Since the operations for converting the number of pixels in the horizontal direction and the vertical direction are the same, only the operation for converting the number of pixels in the horizontal direction will be described.
[0003]
As shown in FIG. 14, when the input image data is composed of a flat portion (h) and a contour portion (j, k), the flat portion (h) and the level changing portion (j, k) are uniformly converted to three times. Therefore, the contour portion is converted into a smooth contour such as j1 and k1.
[0004]
  15 and 16 are diagrams for explaining the detailed operation of the conventional pixel number conversion method. In the figure, p (n) and p (n + 1) are two adjacent pixels of input image data, q (m) is one pixel of output image data, and F (x) is an example of response characteristics of a filter used for pixel number conversion. is there.
[0005]
  When the distance between p (n) and p (n + 1) is 1, and the output image q (m) is located at a distance r from p (n), the output image q (m) is obtained by the following equation.
q (m) = F (r) * p (n) + F (1-r) * p (n + 1)
[0006]
  As shown in FIG. 16, the number of pixels can be converted by performing the above calculation for each pixel (q1 to q7) of the output image data.
[0007]
  In the above description, the case where the linear filter shown in FIG. 15 is used as the response characteristic of the filter is shown. However, the sharpness of the contour (j1, k1) is obtained using the filter having the response characteristic as shown in FIG. May be improved.
[0008]
  Further, as shown in FIG. 18, a method of preparing a plurality of response characteristic filters and switching them according to an image is disclosed in Japanese Patent Laid-Open No. 9-266531.
[0009]
[Problems to be solved by the invention]
  Since the conventional image processing method is configured as described above, when the image is enlarged, a new undershoot (pre-shoot) or overshoot is required to reduce the sharpness of the contour portion or to improve the sharpness. In addition, there is a problem in that when the image is reduced, so-called image quality deterioration of the contour portion occurs such as a pixel of the contour portion is missing.
[0010]
  In addition, there is a problem that the continuity of the image is lost at the filter switching portion by switching the filter according to the image.
[0011]
  The present invention has been made to solve the above-described problems, and it is an object of the present invention to perform image processing while suppressing deterioration in image quality of a contour portion, and more specifically, to change the magnification to an arbitrary magnification.
[0012]
[Means for Solving the Problems]
  An image processing apparatus according to the present invention is an image processing apparatus capable of correcting and outputting the contour of image data representing an input image and converting the number of pixels of the input image into an arbitrary number of pixels,
Features representing changes in image data representing the input imageAs a quantity, the third derivative of the image data, or the difference between the absolute values of the first and second derivativesMeans to seek,
UpThe interpolation pixel that interpolates the input imageThe conversion magnification required for each of the above-mentioned variables for locally changing the density of the interpolation pixel.Exchange rateThe value of theThe, Variables calculated using the above feature quantitiesExchange rate control means,
Interpolation operation means for obtaining an interpolation position of each of the interpolation pixels based on the conversion magnification and calculating pixel data of the interpolation pixels by interpolation calculation at the interpolation position is provided.
[0013]
  An image processing method according to the present invention is an image processing method capable of correcting and outputting a contour of image data representing an input image and converting the number of pixels of the input image into an arbitrary number of pixels,
Features representing changes in image data representing the input imageAs a quantity, the third derivative of the image data, or the difference between the absolute values of the first and second derivativesThe desired process;
Interpolation pixel that interpolates the input imageA conversion magnification required for each of the above-mentioned conversion magnifications for locally changing the density of the interpolation pixel is calculated using the feature amount.Process,
A step of obtaining an interpolation position of each of the interpolation pixels based on the conversion magnification and calculating pixel data of the interpolation pixel by an interpolation calculation at the interpolation position.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
Embodiment 1 FIG.
  FIG. 1 is an explanatory diagram for explaining an image processing method according to Embodiment 1 of the present invention.
  In the figure, the horizontal axis indicates the horizontal or vertical position of the image, and the vertical axis indicates the level (brightness) of the image data. In the drawing, the upper part shows input image data, and the lower part shows output image data when the input image data is enlarged.
[0015]
  Next, the operation will be described.
  In the conversion of the number of pixels, the conversion of the number of vertical pixels and the conversion of the number of horizontal pixels are realized by the same operation. Therefore, here, the operation of conversion of the number of horizontal pixels will be described.
[0016]
  A change in the level of the image is detected from the input image data, and a flat portion (period a) and a contour portion corresponding to the level change portion (periods b and c. That is, the contour portion includes at least three regions of two periods b and c. (Hereinafter, the level changing portion will be described as the contour portion for the sake of simplicity).
[0017]
  Further, the contour portion has different degrees of change in the level of the image (the rate of change is not constant) and the end portion (both are period b, where period b corresponds to the regions at both ends of the level change portion) and the degree of change in the level of the image. Is determined to be the central part of the contour (period c. The area inside the period b, which is the area at both ends of the level changing part, and the area inside the areas at both ends of the level changing part). In this case, the input image data is original image data, and a contour portion in the original image data is detected from a level change between the original image data (level change detection step). Divided).
[0018]
  The determined flat portion a of the image converts the number of pixels at the conversion magnification n. Here, the conversion magnification n is an arbitrary magnification necessary for image format conversion and enlargement or reduction of an image at an arbitrary magnification. As an example of image format conversion, when converting an image of 640 pixels × 480 lines, which is one of the output formats of a personal computer (PC), to an image of 1024 pixels × 768 lines, the conversion magnification n is 1.6 times It is.
[0019]
  On the other hand, in the contour portions (b, c) of the image, the conversion magnification is controlled according to the change amount of the image. That is, in the contour portion divided into a plurality of regions, a generation condition for generating new image data corresponding to each region (the generation condition here is, for example, a conversion magnification. The conversion magnification is processed. (Also referred to as magnification).
[0020]
  More specifically, the start portion and the end portion b of the contour are converted at a higher magnification than the flat portion a, and the central portion c converts the number of pixels at a lower magnification than the flat portion a (that is, the level changing portion). The processing magnification given to the regions at both ends is made larger than the processing magnification given to the region inside the regions at both ends, and as will be described later, the average of the two processing magnifications is the original image. It is set to be equal to the processing magnification given to the flat portion (level flat portion) in the data).
[0021]
  FIG. 2 is a diagram showing a case where input image data is reduced. The flat portion a of the image is reduced at a constant conversion magnification, the start and end portions b of the contour are converted at a higher magnification than the flat portion a, and the central portion c of the contour is converted at a lower magnification than the flat portion a. (The contour portion is divided into a plurality of areas, and the generation conditions corresponding to the respective areas are different (or the generation conditions are different). Of the original image data corresponding to the contour portion based on the detection result obtained by the level change detection process in the above-described enlargement or reduction of the image. A generation condition process for generating a generation condition when generating new image data corresponding to the level change (contour) part, or the original image data corresponding to the level change (contour) part based on the generation condition Level change image data generating step of generating a new image data corresponding to the (outline) portion is included).
[0022]
  The number of pixels can be converted by performing the above operation in the horizontal and vertical directions of the image.
[0023]
  Note that the pixel number conversion in the horizontal direction and the pixel number conversion in the vertical direction can be performed sequentially or simultaneously.
[0024]
  Further, the horizontal conversion magnification and the vertical conversion magnification may be different from each other.
[0025]
  The operation described above is described in the horizontal direction on the display screen. However, if the same operation is performed in the vertical direction, vertical image processing can be realized.
[0026]
  In addition, when a contour portion is detected in the horizontal direction (vertical direction) on the display screen, a new image data generation condition in the horizontal direction (vertical direction when the contour portion is detected in the vertical direction) is output. However, when detecting a level change between adjacent original image data corresponding to the horizontal direction (vertical direction) when performing contour detection, the processing is simplified and the apparatus can be simplified.
[0027]
  Hereinafter, more detailed embodiments will be described with reference to the drawings.
  FIG. 3 is a diagram illustrating the image display apparatus according to Embodiment 1 of the present invention. In the figure, 1 is an image signal input terminal, 2 is a synchronization signal input terminal, 3 is A / D conversion means, 4 is image adjustment means, 5 is memory means for storing input original image data, and 6 is a pixel. A number converter, 7 is an image adjustment means, 8 is a D / A conversion means, 9 is a display means, and 10 is a control means.
[0028]
  An image signal and a synchronization signal are input to the input terminals 1 and 2. The control means 10 generates a sampling clock having a predetermined frequency and a clock and timing signal for controlling the image adjusting means 4 and the following with reference to the synchronization signal input to the input terminal 2. The A / D conversion means 3 samples the image signal input to the input terminal 1 with the sampling clock output from the control means 10 and converts it into digital image data. The image data converted by the A / D conversion unit 3 is input to the image adjustment unit 4 and desired image adjustment is performed.
[0029]
  Here, the desired image adjustment is, for example, when processing the memory means 5 and later as data of three primary colors, when a luminance signal and a color signal are input as an input image signal, or when a composite signal is input. Conversion from luminance signal and color signal to data of three primary colors, conversion from composite signal to data of three primary colors, or vice versa, signals of three primary colors are input, and the memory means 5 and later are processed with luminance signals and color signals. In this case, conversion from data of the three primary colors to data in a format suitable for processing can be considered. Furthermore, arbitrary image adjustment independent of pixel number conversion, such as adjustment of brightness and contrast, may be performed.
[0030]
  The image data processed by the image adjusting means 4 is temporarily stored in the memory means 5 (storage process). Here, the memory means 5 has a capacity capable of storing pixels necessary for the subsequent pixel number conversion (at least two lines or more).
[0031]
  Next, the image data is read out from the memory unit 5 at a predetermined timing under the control of the control unit 10, but may not necessarily be at the same timing as the sampling clock, and any arbitrary necessary for controlling the display unit 9. Can be read by frequency.
[0032]
  The image data Pi read from the memory means 5 is input to the pixel number converter 6 and, as described above, in the contour portion of the image according to the amount of change in the image level (a plurality of regions in the contour portion). Corresponding to each, conversion of the number of pixels is performed while controlling the conversion magnification, and the converted image data Po is output (that is, a digital linear filter as generally known here). Is used to convert the number of pixels).
[0033]
  The image data output from the pixel number converter 6 is input to the image adjusting means 7 and desired image adjustment is performed. Here, the desired image adjustment refers to various processes and corrections independent of pixel number conversion such as adjustment of brightness, contrast and saturation and gradation control, and conversion of a signal format for input to the display means 9 Etc. are included.
[0034]
  The image data processed by the image adjustment unit 7 is input to the D / A conversion unit 8 and converted into an analog image signal. The image signal converted by the D / A conversion unit 8 is input to the display unit 9 and displayed at a predetermined timing under the control of the control unit 10.
[0035]
  Although FIG. 3 shows a configuration in which the D / A conversion means 8 converts it into an analog image signal and inputs it to the display means 9, but the display means 9 can directly input and display digital image data. Can omit the D / A conversion means 8 (the operation of the display means 9 corresponds to the display process).
[0036]
  Next, a more detailed operation of the pixel number converter 6 will be described.
  Here, the pixel number converter 6 may be configured to perform the pixel number conversion independently in each of the horizontal and vertical directions, but here the case where the pixel number conversion is performed in both the vertical direction and the horizontal direction. explain.
[0037]
  FIG. 4 is a diagram showing a detailed configuration of the pixel number converter 6 according to the first embodiment of the present invention.
  In the figure, 11 is a vertical pixel number conversion means, 12 is a vertical high frequency component detection means, 13 is a vertical conversion magnification control means, 14 is a horizontal pixel number conversion means, 15 is a horizontal high frequency component detection means, and 16 is a horizontal conversion magnification. It is a control means.
[0038]
  The vertical high-frequency component detection means 12 outputs the vertical first-order differential result vd1 and the third-order differential result vd3 as the high-frequency component (level change amount) in the vertical direction of the image data Pi. Here, the image data Pi is image data read from the memory means 5 shown in FIG. 3, and is composed of a plurality of pixels required by the vertical pixel number converting means 11 and the vertical high frequency component detecting means 12. .
[0039]
  The vertical conversion magnification control means 13 obtains a vertical conversion magnification vc1 from the vertical high-frequency components vd1 and vd3 output from the vertical high-frequency component detection means 12, and outputs the vertical conversion magnification vc1 to the vertical pixel number conversion means 11. The vertical pixel number conversion means 11 converts the number of pixels in the vertical direction of the input image based on the vertical conversion magnification vc1 and outputs the conversion result Pv.
[0040]
  Next, the horizontal high-frequency component detection means 15 uses the horizontal first-order differential results hd1 and tertiary as the horizontal high-frequency component (level change amount) of the image data Pv output from the vertical pixel number conversion means 11. The differentiation result hd3 is output.
[0041]
  The horizontal conversion magnification control means 16 obtains a horizontal conversion magnification hc1 from the horizontal high-frequency components hd1 and hd3 output from the horizontal high-frequency component detection means 15, and outputs the horizontal conversion magnification hc1 to the horizontal pixel number conversion means 14. The horizontal pixel number conversion means 14 converts the number of pixels in the horizontal direction of the input image based on the horizontal conversion magnification hc1, and outputs the conversion result Po.
[0042]
  In the configuration shown in FIG. 4, the vertical high-frequency component detection means 12 and the horizontal high-frequency component detection means 15 function as contour detection means, and the vertical conversion magnification control means 13 and horizontal conversion magnification control means 16 function as generation condition generation means. The vertical pixel number conversion means 11 and the horizontal pixel number conversion means 14 function as image data generation means (of course, a configuration for either vertical or horizontal data processing, for example, generation of new image data in the vertical direction) In FIG. 5, the image processing system may be configured only by the vertical high-frequency component detection means 12, the vertical conversion magnification control means 13, and the vertical pixel number conversion means 11, and the same applies to the horizontal direction).
[0043]
  Although described with reference to the configuration shown in FIGS. 3 and 4, a low-pass filter (low-pass filter) for removing noise immediately before the vertical high-frequency component detection unit 12 with respect to the output Pi from the memory unit 5. A filter (LPF) may be provided. In this case, the output Pi is inputted as it is to the vertical pixel number converting means 11, and the output of the LPF provided immediately before is inputted to the vertical high frequency component detecting means 12. In this way, noise existing on the output Pi is not processed as a contour portion, and only the contour portion can be processed accurately (in this case, the vertical high-frequency component detection means 12 and the LPF). Contour detection means is configured).
[0044]
  Of course, such an LPF can also be arranged immediately before the horizontal high-frequency component detection means 15, and the same effect as the vertical data processing can be obtained in the horizontal direction (in this case, the horizontal high-frequency component detection means 15). The contour detection means is constituted by the band component detection means 15 and the LPF). Of course, it is possible to employ these configurations in one of the vertical direction and the horizontal direction, or in both directions.
[0045]
  In the above description, the vertical high-frequency component detecting means 12 in the vertical direction or the LPF provided in the preceding stage is described. However, the preceding stage of the vertical conversion magnification control means 13 is a band pass filter (BPF). Can be expected to have the same effect (it goes without saying that the same effect can be obtained by adopting the same configuration in the horizontal direction). In this case, the contour detecting means is constituted by the vertical high frequency component detecting means 12 and the BPF (or the horizontal high frequency component detecting means 15 and the BPF).
[0046]
  Further, the contour detecting means is constituted by the vertical high frequency component detecting means 12 (or horizontal high frequency component detecting means 15) or the vertical high frequency component detecting means 12 and LPF (or horizontal high frequency component detecting means 15 and LPF). A coring means for providing a dead zone is provided after the vertical high-frequency component detection means 12 (or horizontal high-frequency component detection means 15) or before the vertical conversion magnification control means 13 (or horizontal conversion magnification control means 16). This can also provide the same effects as described above.
  In the above description, the configuration example in which the pixel number conversion is performed in both the vertical and horizontal directions has been described, but the operation of the pixel number conversion will be described below. In order to facilitate understanding, the horizontal pixel number conversion operation will be described.
[0047]
  FIG. 5 is a diagram for explaining in detail the operation of converting the number of horizontal pixels.
  In FIG. 5, hd1 and hd3 indicate the first-order differential result and the third-order differential result in the horizontal direction corresponding to the image data Pv. Here, the horizontal conversion magnification hc1 is
hc1 = n + k × hd1 × hd3
As shown, the result obtained by multiplying the first derivative result hd1 and the third derivative result hd3 by an arbitrary number k is added to the conversion magnification n. As a result, the number of pixels is converted at the conversion magnification n in the period a, the magnification higher than the conversion magnification n in the period b, and the magnification lower than the conversion magnification n in the period c. However, since the average value of the horizontal conversion magnification hc1 in one line is the conversion magnification n as shown by the following expression, the conversion magnification of the image locally rises and falls, but the conversion magnification of the entire image is n (horizontal The number of pixels in the direction is the original number of pixels in the same direction × n).
  AVE (hc1) = n
  However, AVE (x) represents the average value of one line of the conversion magnification x.
[0048]
  In the above description, the case where the horizontal conversion magnification hc1 is multiplied by an arbitrary number k multiplied by the product of the first derivative result hd1 and the third derivative result hd3 is added to the conversion magnification n. However, instead of multiplying an arbitrary number k by limiting the maximum and minimum values to the result of multiplying an arbitrary number k, non-linear conversion corresponding to the horizontal position (or vertical position) of the image is performed. Thus, the conversion magnification having an arbitrary characteristic in the contour portion can be freely controlled.
[0049]
  In the vertical direction, the operation is the same as the above-described horizontal pixel number conversion. In this vertical case, the average value of the conversion magnification in the vertical direction is n (the number of pixels in the vertical direction is the same as that in the original direction). (The number of pixels × n) (when the vertical conversion magnification is n. Of course, the horizontal and vertical conversion magnifications can be set independently).
[0050]
  FIG. 6 is an explanatory diagram for explaining the operation of converting the number of pixels in the contour portion of FIG. 5 in more detail.
[0051]
  A part of pixels (p1, p2, p3) of the image data Pv is converted into q11 to q17 by the above operation. In the period b, the pixel density is higher than in the period a, and in the period c, the pixel density is lower than in the period a. As a result (q11 to q20) are displayed on the display means 9 at equal intervals as shown in s11 to s20, the portion having a high pixel density is enlarged at a magnification higher than the conversion magnification n, and the pixel density is low. In the portion, the image is enlarged at a magnification lower than the conversion magnification n.
[0052]
  More specifically, the distance between q11 and q12 is represented by the reciprocal of the conversion magnification at q12, where the distance between p1 and p2 (pixel interval before processing) is 1. The distance from q11 to q13 may be obtained by adding the reciprocal of the conversion magnification at q13 to the distance from q11 to q12. In this way, the position of each pixel is obtained by cumulatively adding the reciprocal of the conversion magnification hc1 output from the horizontal conversion magnification control means 16 in the horizontal direction.
[0053]
  More specifically, from the result of this cumulative addition, pixels required for processing (for example, addresses of memory means storing pixels such as p1, p2, and p3) and filter coefficients (filter coefficient numbers (filter coefficients are In the case of the lookup table, the number (address)) and the filter coefficient itself can be obtained.
[0054]
  Through the above operation, some of the pixels (p1, p2, p3) of the image data Pv are converted as q11 to q17. In the period b, the pixel density is higher than in the period a, and in the period c, the pixel density is lower than in the period a. More specifically, since the results (q11 to q20) are displayed on the display means 9 at equal intervals as indicated by s11 to s20, a portion with a high pixel density is converted at a magnification higher than the conversion magnification n. In a portion where the pixel density is low, the pixel is converted (enlarged) at a magnification lower than the conversion magnification n.
[0055]
  In the above description of the operation, the case of enlarging the image has been described. However, the same applies to the case of reducing the image. In the period b, conversion (reduction) is performed at a higher magnification than the conversion magnification n. Conversion (reduction) is performed at a lower magnification.
[0056]
  Therefore, when the conversion magnification n is larger than 1, the image is enlarged without impairing the sharpness of the contour portion, and when it is smaller than 1, the image is shrunk so that the lack of the contour portion image is reduced. That is, there is an effect that the contour information of the input image is preserved in both enlargement and reduction.
[0057]
  Also, by taking an arbitrary number k large, the sharpness of the contour of the input image can be increased, and the sharpness of the contour portion can be controlled by the arbitrary number k.
[0058]
  When the conversion magnification n is 1, the entire image is not enlarged or reduced, and only the sharpness of the contour portion is controlled.
[0059]
  In addition, by setting the conversion magnification n and an arbitrary number k independently in the vertical direction and the horizontal direction, the horizontal conversion magnification and the horizontal outline sharpness, and the vertical conversion magnification and the vertical outline Sharpness can be controlled independently.
[0060]
  For example, by setting the vertical conversion magnification n to 2 and the horizontal conversion magnification n to 1, conversion from an interlaced image to a non-interlaced image (scanning line interpolation) can be performed. And the vertical contour can be independently controlled to a desired sharpness.
[0061]
  In the above description of the operation, the case where the vertical pixel number conversion and the horizontal pixel number conversion operation are sequentially performed as the pixel number conversion operation has been described. However, after the horizontal pixel number conversion, the vertical direction conversion is performed. The same effect can be obtained by converting the number of pixels.
[0062]
  Further, the same effect can be obtained even if the vertical pixel number conversion and the horizontal pixel number conversion are simultaneously performed.
[0063]
  In the above description of the operation, the case where a linear filter is used as the pixel number converter used for the conversion of the number of pixels has been described. However, an arbitrary filter such as a nonlinear filter can be used.
[0064]
  FIG. 7 is a diagram showing a configuration when the pixel number converter 6 performs the pixel number conversion in the vertical direction and the horizontal direction at the same time. In the figure, 17 is a pixel number conversion means, 18 is a high frequency component detection means, and 19 is a conversion magnification control means. The high frequency component detection means 18 outputs the primary differential result d1 and the tertiary differential result d3 as the vertical and horizontal change amounts of the image data Pi. The conversion magnification control unit 19 determines the conversion magnification c1 based on the primary differentiation result d1 and the secondary differentiation result d2, and outputs the conversion magnification c1 to the pixel number conversion unit 17. The pixel number conversion means 17 converts the number of pixels in the vertical direction and the horizontal direction two-dimensionally simultaneously based on the conversion magnification c1.
[0065]
  The relationship between the high frequency component detection results d1 and d3 and the conversion magnification c1 is similar to the relationship between the horizontal conversion magnification hc1 obtained using the first and third differential results hd1 and hd3 in FIG. Detailed description is omitted.
[0066]
Embodiment 2. FIG.
  In the first embodiment, the configuration in which the conversion magnification is controlled using the primary differential result and the tertiary differential result as the high-frequency component (level change amount) of the image has been described. However, the primary differential result and the secondary differential are described. By controlling the conversion magnification using the differential result, the configuration can be simplified (the third differential result is obtained by differentiating the second differential result, so one differential means can be omitted) it can).
[0067]
  Hereinafter, a more detailed operation of the pixel number converter 6 will be described (other configurations are the same as those described in the first embodiment).
  Here, the pixel number converter 6 may be configured to perform the pixel number conversion independently in each of the horizontal and vertical directions, but here the case where the pixel number conversion is performed in both the vertical direction and the horizontal direction. explain.
[0068]
  FIG. 8 is a diagram illustrating the pixel number converter 6 according to the second embodiment.
  In the figure, 20 is a vertical high frequency component detection means, 21 is a vertical conversion magnification control means, 22 is a horizontal high frequency component detection means, and 23 is a horizontal conversion magnification control means.
[0069]
  The vertical high frequency component detection means 20 outputs the primary differential result vd1 and the secondary differential result vd2 of pixels adjacent in the vertical direction as the high frequency component (level change amount) in the vertical direction of the image data Pi. The vertical conversion magnification control means 21 obtains the vertical conversion magnification vc2 from the first-order differential result vd1 and the second-order differential result vd2 output from the vertical high-frequency component detection means 20, and supplies the vertical pixel number conversion means 11 with it. Output. The vertical pixel number conversion means 11 converts the number of pixels in the vertical direction of the input image based on the vertical conversion magnification vc2, and outputs the conversion result Pv.
[0070]
  Next, the horizontal high-frequency component detection unit 22 uses the first-order differential result hd1 of the pixels adjacent in the horizontal direction as the high-frequency component (level change amount) in the horizontal direction of the image data Pv output from the vertical pixel number conversion unit 11. And the second derivative result hd2. The horizontal conversion magnification control means 23 obtains a horizontal conversion magnification hc2 from the horizontal primary differential result hd1 and the secondary differential result hd2 output from the horizontal high frequency component detection means 22, and outputs the horizontal conversion magnification hc2 to the horizontal pixel number conversion means 14. To do. The horizontal pixel number conversion means 14 converts the number of pixels in the horizontal direction of the input image based on the horizontal conversion magnification hc2, and outputs the conversion result Po.
[0071]
  In the above description, the configuration example in which the pixel number conversion is performed in both the vertical and horizontal directions has been described, but the operation of the pixel number conversion will be described below. In order to facilitate understanding, the horizontal pixel number conversion operation will be described.
[0072]
  FIG. 9 is a diagram showing the configuration of the horizontal conversion magnification control means 23.
  In the figure, 24 and 25 are absolute value conversion means, and 26 is a signal subtraction means for subtracting the output of the absolute value conversion means 24 from the output of the absolute value conversion means 25.
[0073]
  FIG. 10 is a diagram for explaining the operation of converting the number of horizontal pixels in detail.
  In FIG. 10, hd1 and hd2 indicate the first-order differential result and the second-order differential result in the horizontal direction corresponding to the image data Pv. Here, the horizontal conversion magnification hc2 is
hc2 = n + k × (abs (hd2) −abs (hd1))
As shown by the following formula, the result obtained by multiplying the conversion magnification n by the absolute value abs (hd2) of the first derivative result hd1 from the absolute value abs (hd2) of the second derivative result hd2 is multiplied by an arbitrary number k. It is a thing. Note that the conversion magnification n and the arbitrary number k are the same as those in the first embodiment (contour detection step. The contour portion detected here is divided into a plurality of regions).
  Here, abs (x) indicates the absolute value of the differentiation result x.
[0074]
  As a result, the number of pixels is converted with the conversion magnification n in the period a of hc2, the magnification higher than the conversion magnification n in the period b, and the magnification lower than the conversion magnification n in the period c. That is, a generation condition (for example, the generation condition here is a conversion magnification) when generating new image data corresponding to each area of the contour divided into a plurality of areas corresponds to each area. Are set (here, they are set differently). However, since the average value of the horizontal conversion magnification hc2 in one line is the conversion magnification n as shown by the following expression, the conversion magnification of the image is locally increased or decreased, but the conversion magnification of the entire image is n (horizontal direction). Is the original number of pixels in the same direction × n).
  AVE (hc1) = n
  However, AVE (x) represents the average value of one line of the conversion magnification x.
[0075]
  Here, a generation condition step for generating a generation condition when generating new image data corresponding to the contour portion from the original image data (input image data) corresponding to the contour portion based on the detection result obtained by the contour detection step And an image data generation step of generating new image data corresponding to the contour portion from the original image data corresponding to the contour portion based on the generation condition.
[0076]
  The above-described operation is described in the horizontal direction on the display screen. However, if the same operation is performed in the vertical direction, vertical image processing can be realized. The average value of conversion magnification is n (the number of pixels in the vertical direction is the original number of pixels in the same direction × n) (when the conversion magnification in the vertical direction is n. Of course, the conversion magnifications in the horizontal direction and the vertical direction are independent of each other. Can be set).
[0077]
  Since other operations are the same as those in the first embodiment, the description thereof is omitted.
[0078]
  FIG. 11 is a diagram showing a configuration in the case of simultaneously converting the number of pixels in the vertical direction and the horizontal direction. In the figure, reference numeral 27 denotes high-frequency component detection means, and 28 denotes conversion magnification control means. The high frequency component detecting means 27 outputs a primary differential result d1 and a secondary differential result d2 as the vertical and horizontal change amounts of the image data Pi. The conversion magnification control unit 27 determines the conversion magnification c2 based on the primary differentiation result d1 and the secondary differentiation result d2, and outputs the conversion magnification c2 to the pixel number conversion unit 17. The pixel number conversion means 17 converts the number of pixels in the vertical direction and the horizontal direction at the same time (two-dimensional conversion) based on the conversion magnification c2.
[0079]
  The relationship between the high frequency component detection results d1 and d2 and the conversion magnification c2 is the same as the relationship between the primary and secondary differential results hd1 and hd2 and the horizontal conversion magnification hc2 in FIGS. The detailed explanation is omitted.
[0080]
Embodiment 3 FIG.
  In the first and second embodiments, the case where the input image signal is an analog signal has been described. However, the present invention is not limited to this, and digital image data may be input.
[0081]
  FIG. 12 is a diagram showing an image display device according to Embodiment 3 of the present invention.
  In the figure, 29 is an input terminal for digital image data, 30 is a display means capable of directly inputting digital data, and 31 is a control means (other configurations are the same as those in the first and second embodiments).
[0082]
  Next, the operation will be described.
  Digital image data is input to the input terminal 29, and the image data input to the input terminal 29 is input to the image adjustment unit 4. A synchronization signal is input to the input terminal 2, and the synchronization signal input to the input terminal 2 is input to the control means 31.
[0083]
  The image adjustment unit 4, the memory unit 5, the pixel number converter 6, and the image adjustment unit 7 are controlled by the control unit 31 to perform the conversion of the number of pixels and other operations in the same manner as in the first and second embodiments. Perform image processing. The image data output from the image adjusting unit 7 is directly input to the display unit 30 and displayed at a predetermined timing under the control of the control unit 31. Since other detailed operations are the same as those in the first and second embodiments, a description thereof will be omitted.
[0084]
  In the description of the operation in the third embodiment, the display means 30 capable of directly inputting digital image data has been described. However, the D / A conversion means 8 shown in the first embodiment and the display means 30 are used instead of the display means 30. It can also be configured using the display means 9.
[0085]
Embodiment 4 FIG.
  In the first to third embodiments, the configuration in which the number of pixels is converted by hardware has been described. However, the number of pixels can also be converted by software. FIG. 13 is a flowchart for explaining the pixel number conversion operation (image processing method / image display method) by software processing (of course, software and hardware may be mixed).
[0086]
  Next, the operation will be described.
  Here, the pixel number conversion may be configured such that the pixel number conversion is performed independently in each of the horizontal and vertical directions, but here the case where the pixel number conversion is performed in both the vertical and horizontal directions. explain.
[0087]
  (Start of data generation operation in the vertical direction according to the flow of A in the figure) In the data extraction unit, the calculation of the vertical high-frequency component for the pixel of interest and the filter operation from the image data for converting the number of pixels (corresponding to Pi in FIG. 4) A plurality of necessary pixel data is extracted.
[0088]
  In the vertical high frequency component calculation unit, the first-order differential result in the vertical direction (corresponding to vd1 in FIG. 4) and the third-order differential result in the vertical direction (corresponding to vd3 in FIG. 4) from the plurality of pixel data extracted by the data extraction unit. ) (This corresponds to the level change detection step).
[0089]
  In the vertical conversion magnification calculation unit, the vertical conversion magnification (with respect to the pixel of interest) is calculated from the primary differentiation result and the secondary differentiation result calculated by the vertical high frequency component calculation unit and the conversion magnification of the entire image (corresponding to n in FIG. 5). (Corresponding to vc1 in FIG. 4) is calculated (generation condition generation step).
[0090]
  The filter calculation unit performs a filter calculation from the conversion magnification calculated by the vertical conversion magnification calculation unit and the plurality of pixel data extracted by the data extraction unit, and stores the calculation result (image data generation step).
[0091]
  The above operation is repeated until the target pixel reaches the end of the image. Here, the end of the image indicates the right end of the image when computing from the left side of the image.
[0092]
  When the target pixel reaches the end of the image, the target pixel is moved to the next line and the above calculation is performed. By performing this operation on all the pixels, the conversion of the number of pixels in the vertical direction is completed (end of the operation according to the flow A in the figure).
[0093]
  (Start of horizontal data generation operation according to the flow B in the figure) In the next data extraction unit, the horizontal high frequency for the pixel of interest is calculated from the image data (corresponding to Pv in FIG. 4) in which the number of pixels in the vertical direction is converted. A plurality of pixel data necessary for component calculation and filter calculation are extracted. In the horizontal high-frequency component calculation unit, the first-order differential result in the horizontal direction (corresponding to hd1 in FIG. 4) and the third-order differential result in the horizontal direction (corresponding to hd3 in FIG. 4) from the plurality of pixel data extracted by the data extraction unit. ) Is calculated. The horizontal conversion magnification calculator calculates the horizontal conversion magnification (with respect to the pixel of interest) from the first and third differential results calculated by the horizontal high-frequency component calculator and the conversion magnification of the entire image (corresponding to n in FIG. 5). (Corresponding to hc1 in FIG. 4). The filter calculation unit performs a filter calculation from the conversion magnification calculated by the horizontal conversion magnification calculation unit (a generation condition when generating new image data) and a plurality of pixel data extracted by the data extraction unit, and the calculation result Save.
[0094]
  Next, the above operation is repeated until the target pixel reaches the end of the image.
[0095]
  When the target pixel reaches the end of the image, the target pixel is moved to the next line and the above calculation is performed. By performing this operation on all the target pixels, the conversion of the number of pixels is completed (end of the operation by the flow B in the figure).
[0096]
  In the above description of the operation, the case of converting the number of pixels in the horizontal direction after converting the number of pixels in the vertical direction has been described. However, the number of pixels in the vertical direction may be converted after converting the number of pixels in the horizontal direction. Good (that is, after the operation according to the flow B in the figure is first performed, the operation according to the flow A in the figure may be performed). Further, as described above, either the operation according to the flow A in the drawing or the operation according to the flow B in the drawing may be performed.
[0097]
  In the above description of the operation, the case of calculating the pixel of interest in the order from the left to the right and the top to the bottom of the image when converting the number of vertical and horizontal pixels is not limited to this. The same result can be obtained by calculating from.
[0098]
  Note that the average value n of the conversion magnifications in one line described above (when the conversion magnification in either the vertical or horizontal direction is n. Of course, the conversion magnifications in the vertical and horizontal directions can be set independently. Can be established in both the vertical and horizontal directions (when the conversion magnification is n in both directions) or in either direction.
[0099]
【The invention's effect】
  Image processing apparatus according to the present inventionAccording to the image processing method, the amount of change in magnification for adjusting the density of the interpolation pixel according to the change in the image data is obtained, and the pixel data of the interpolation pixel is calculated using the amount of change in magnification. The input image can be converted into an arbitrary number of pixels without impairing the image quality.
[Brief description of the drawings]
FIG. 1 is a diagram showing an image processing operation in Embodiment 1. FIG.
FIG. 2 is a diagram illustrating an image processing operation in the first embodiment.
3 is a diagram illustrating a configuration of an image display device according to Embodiment 1. FIG.
4 is a diagram illustrating a configuration of an image processing apparatus according to Embodiment 1. FIG.
FIG. 5 is a diagram illustrating an image processing operation in the first embodiment.
FIG. 6 is a diagram showing an image processing operation in the first embodiment.
7 is a diagram showing another configuration of the image processing apparatus according to Embodiment 1. FIG.
FIG. 8 is a diagram illustrating a configuration of an image processing apparatus according to a second embodiment.
9 is a diagram illustrating a configuration of an image processing apparatus according to Embodiment 2. FIG.
FIG. 10 is a diagram illustrating an image processing operation in the second embodiment.
FIG. 11 is a diagram illustrating another configuration of the image processing apparatus according to the second embodiment.
12 is a diagram illustrating a configuration of an image display device in Embodiment 3. FIG.
FIG. 13 is a flowchart of an image processing operation in the fourth embodiment.
FIG. 14 is a diagram illustrating a conventional image processing operation.
FIG. 15 is a diagram illustrating a conventional image processing operation.
FIG. 16 is a diagram illustrating a conventional image processing operation.
FIG. 17 is a diagram illustrating an example of response characteristics of a conventional image processing unit.
FIG. 18 is a diagram illustrating an example of response characteristics of a conventional image processing unit.
[Explanation of symbols]
  1 input terminal, 2 input terminal, 3 A / D conversion means, 4 image adjustment means, 5 memory means, 6 pixel number converter, 7 image adjustment means, 8 D / A conversion means, 9 display means, 10 control means, 11 vertical pixel number conversion means, 12 vertical high frequency component detection means, 13 vertical conversion magnification control means, 14 horizontal pixel number conversion means, 15 horizontal high frequency component detection means, 16 horizontal conversion magnification control means, 17 pixel number conversion means, 18 High-frequency component detection means, 19 Conversion magnification control means, 20 Vertical high-frequency component detection means, 21 Vertical conversion magnification control means, 22 Horizontal high-frequency component detection means, 23 Horizontal conversion magnification control means, 24 and 25 Absolute value conversion means , 26 signal subtracting means, 27 high frequency component detecting means, 28 conversion magnification control means, 29 input terminal, 30 display means, 31 control means.

Claims (7)

An image processing apparatus capable of correcting and outputting the contour of image data representing an input image and converting the number of pixels of the input image into an arbitrary number of pixels,
Means for obtaining a third derivative of the image data, or a difference between absolute values of the first and second derivatives as a feature quantity representing a change in image data representing the input image;
A conversion ratio obtained for each of the interpolated pixel to be interpolated on the entry force image, varying換倍rate the value of the variable換倍rate for locally changing the density of the interpolated pixel is calculated by using the feature amount Control means;
An image processing apparatus comprising: an interpolation calculation unit that obtains an interpolation position of each of the interpolation pixels based on the conversion magnification and calculates pixel data of the interpolation pixel by an interpolation calculation at the interpolation position.   The image processing apparatus according to claim 1, wherein the conversion magnification control unit obtains the conversion magnification by adjusting a reference conversion magnification of the interpolation pixel that is arbitrarily set using the adjustment amount. Interpolation calculation means, by accumulating the reciprocal of the conversion ratio in the horizontal and / or vertical direction, to claim 1 or 2, characterized in that to calculate the interpolation position in the horizontal and / or vertical direction of the interpolation pixel The image processing apparatus described. The image display apparatus comprising the image processing apparatus according to any one of claims 1-3. An image processing method capable of correcting and outputting the contour of image data representing an input image and converting the number of pixels of the input image into an arbitrary number of pixels,
Obtaining a third derivative of the image data, or a difference between absolute values of the first and second derivatives, as a feature amount representing a change in image data representing the input image;
A conversion magnification obtained for each of the interpolation pixels for interpolating the input image, the step of calculating the value of the conversion magnification for locally changing the density of the interpolation pixels using the feature amount ;
An image processing method comprising: calculating an interpolation position of each of the interpolation pixels based on the conversion magnification, and calculating pixel data of the interpolation pixel by an interpolation calculation at the interpolation position.   The image processing method according to claim 5, wherein the conversion magnification is obtained by adjusting a reference conversion magnification of the interpolation pixel using the adjustment amount.   7. The image processing according to claim 5, wherein an interpolation position in the horizontal and / or vertical direction of the interpolated pixel is calculated by accumulatively adding the reciprocal of the conversion magnification in the horizontal and / or vertical direction. Method.

Images