JP4641957B2 - Image processing apparatus and image processing program - Google Patents

Description

  The present invention relates to an image processing apparatus and an image processing program, and is particularly suitable for use in an apparatus or a program that imparts an image effect by performing predetermined pixel value conversion processing on image data.

  2. Description of the Related Art Conventionally, an apparatus for providing an image effect by performing predetermined pixel value conversion processing on image data has been provided. Many of this type of image processing apparatus performs a predetermined calculation using the pixel value of a target pixel to be converted and the pixel values of peripheral pixels located around the target pixel, and uses the resulting value as the target pixel. A new pixel value is used as a new pixel value. Among them, there is one that obtains a difference between a pixel value of a target pixel and pixel values of surrounding pixels and obtains a new pixel value using the difference value.

For example, the hues of the target pixel and its surrounding pixels are compared, and when the difference between the two hues exceeds a predetermined value, it is determined that the target pixel is a false color, and the hue of the target pixel is controlled by the surroundings. Some of them are replaced with typical hues (see, for example, Patent Document 1). In Patent Document 1, the difference between the hue of the pixel of interest and the hue of surrounding pixels is obtained, and when the difference is small, the weight is increased, and when the difference is large, the weight is decreased and the weight coefficient for the input image is obtained. It is also described.
JP-A-2005-260675

Also, a high frequency region having a coefficient difference calculating means for calculating a correction coefficient corresponding to the signal level of the luminance signal and a correcting means for correcting the luminance signal based on the correction coefficient, wherein the difference in the luminance signal level between adjacent pixels is large. Then, by setting a narrowband correction coefficient in the correction means, and setting a wideband correction coefficient in the correction means in the low frequency region where the difference in luminance signal level is small, the contour of the image can be enhanced with higher accuracy. Such a device is also provided (see, for example, Patent Document 2).
JP 2000-13642 A

Further, a difference value calculation unit that calculates a difference value between the pixel value of the target pixel and the pixel values of the surrounding pixels, a filter coefficient calculation unit that calculates a filter coefficient for the surrounding pixels based on the calculated difference value, and There is also provided an apparatus for generating a restored image in which mosquito noise generated by block coding is effectively removed by including a filter processing unit that performs a filtering process on the pixel value of the target pixel using the filter coefficient obtained. (For example, see Patent Document 3).
JP 2004-159511 A

Further, the edge characteristic of the input image data is detected, a value k indicating the degree of edge is obtained from the detected edge characteristic, a value d corresponding to the k is obtained by a conversion table, and the obtained value of d is obtained. There is also an apparatus that obtains the component value of each element of the smoothing filter by using (see, for example, Patent Document 4).
Japanese Patent Laid-Open No. 11-185036

  In the conventional techniques such as Patent Documents 1 to 4, the correction coefficient to be multiplied with respect to the input image is made variable according to the difference between the pixel values of the target pixel and the peripheral pixels, for example, corresponding to the contour. Only the portion can be emphasized, or only the portion corresponding to false color or mosquito noise can be blurred.

  However, the conventional technique has a problem that the image quality obtained by the pixel value conversion process using the correction coefficient is not always good. For example, in the conventional technique, the only way to change the strength of the image effect is to change the value of the correction coefficient. For example, in order to increase the degree of contour enhancement, it is necessary to increase the correction coefficient value. However, in this case, the difference in pixel values between the emphasized contour portion and the peripheral portion thereof becomes unnaturally large, and the contour enhancement degree is increased, but the image becomes unsightly. The same is true for the blur effect.

  Further, in the above conventional technique, only the same image effect can be applied to one image, such as contour enhancement for contour enhancement and blurring for blur. Therefore, composite image effects such as suppressing mosquito noise while emphasizing contours, or suppressing discomfort between the contour portion emphasized while emphasizing contours and its peripheral portion cannot be provided. . In other words, the conventional technique can only give a side effect, and there is a problem that the image quality obtained by the image effect is not always good.

  The present invention has been made to solve such a problem, and an image is obtained by pixel value conversion processing using a correction coefficient determined according to a difference in pixel value between a pixel of interest and its surrounding pixels. It is an object to improve the image quality of an image to which an effect is given.

  In order to solve the above-described problem, the image processing apparatus according to the present invention includes a plurality of patterns prepared as a positional relationship between two pixels for which a difference is taken, and stored with different correction values for each pattern. Then, a difference value between the pixels is calculated for each pattern, and based on the calculated difference value, the table information regarding the same pattern as the pattern for which the difference value is obtained is referred to, and a correction value corresponding to the difference value is obtained. I am doing so. Here, the difference value between pixels does not need to be calculated for all patterns.

  According to the present invention configured as described above, different table information is referred to according to the positional relationship between two pixels that have taken a difference, and a correction value is obtained from the table information. The contents of a plurality of table information can be arbitrarily set, and even if the difference value is the same, different correction values can be set for each table. In this way, even if the calculated difference values are the same, the correction values obtained can be different depending on the positional relationship between the two pixels that have taken the difference. Further, the obtained correction value can be different depending on which pattern's positional relationship the difference value is obtained, that is, which pattern's table information is used to obtain the correction value.

  Accordingly, abundant variations of image effects that have not been available can be provided by using table information or a combination thereof. For example, a new image effect such as contour enhancement + blurring can be given. In addition, since an appropriate correction value can be set according to the positional relationship between the pixels taking the difference, the image quality of the image obtained by the pixel value conversion process using the correction value can also be improved.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a functional block diagram illustrating a configuration example of the image processing apparatus according to the present embodiment. As shown in FIG. 1, the image processing apparatus according to this embodiment includes a difference value calculation unit 1, a table information storage unit 2, a correction value acquisition unit 3, and a pixel value conversion unit 4. .

The difference value calculation unit 1 sets a processing area composed of a range of m pixels vertically and n pixels horizontally (m and n are integers of 1 or more and at least one is 2 or more) for a part of the input image. Then, within the processing area, a difference between pixel values (for example, luminance values) is calculated for two pixels having a predetermined positional relationship. Here, the difference value calculation unit 1 includes a plurality of patterns of difference calculation units 1 ₋₁ , 1 ₋₂ ,..., 1 _−k (k is an integer of 2 or more), and a plurality of prepared positional relationships. For each pattern, a difference value between two pixels is calculated.

  2-4 is a figure which shows the example of the pattern considered as a positional relationship of the two pixels which take a difference. 2 to 4, A to I each represent a pixel or a pixel value thereof. Here, an example is shown in which a range of 9 pixels of 3 vertical pixels × 3 horizontal pixels is used as a processing region. An example in the case of k = 3 is shown.

In the first pattern shown in FIG. 2, the difference value calculation unit 1 includes a target pixel E set at the center in the processing region and peripheral pixels (adjacent pixels) positioned in the vertical and horizontal directions with respect to the target pixel E. ) Calculate difference values from B, D, F, and H, respectively. That is, the difference computing section _{1 -1} of the first pattern, four difference values be shown below to calculate de, fe, the he.
be = BE, de = DE, fe = FE, he = HE

In the second pattern shown in FIG. 3, the difference value calculation unit 1 calculates the difference value between two peripheral pixels B and H positioned in the vertical direction with the target pixel E set at the center in the processing region, and the target The difference value between the two peripheral pixels D and F located in the horizontal direction across the pixel E is calculated. That is, the difference computing section _{1 -2} second pattern, two difference values bh shown below to calculate the df. Note that || indicates an absolute value.
bh = | B−H |, df = | D−F |

In the third pattern shown in FIG. 4, the difference value calculation unit 1 includes a pixel of interest E set at the center in the processing region and peripheral pixels A, C, G, A difference value from I is calculated. That is, the third pattern difference computing section _{1 -3} of the four difference values ae shown below to calculate ce, ge, and ie.
ae = AE, ce = CE, ge = GE, ie = IE

Returning to FIG. The table information storage unit 2 stores in advance correction values corresponding to the difference values between two pixels as table information. For example, a hard disk, a semiconductor memory, a CD (Compact Disc), a DVD (Digital Versatile Disk), It is constituted by a recording medium such as a ROM or a RAM. The table information storage unit 2 stores a plurality of patterns of tables 2 ₋₁ , 2 ₋₂ ,..., 2- _k storing different correction values for the plurality of patterns illustrated in FIGS. .

The correction value acquisition unit 3 refers to the table information stored in the table information storage unit 2 based on the difference value calculated by the difference value calculation unit 1 and obtains a correction value corresponding to the difference value. Here, the correction value acquisition unit 3 includes a plurality of correction value acquisition units 3 ₋₁ , 3 ₋₂ ,..., 3- _k corresponding to the plurality of patterns described above. The correction value acquisition units 3 ₋₁ , 3 ₋₂ ,..., 3- _k of each pattern are difference values calculated for each pattern by the difference value calculation units 1 ₋₁ , 1 _{−2 ,.} The correction value corresponding to the difference value is obtained with reference to the correction value tables 2 ₋₁ , 2 _{−2 ,.}

For example, the correction value acquisition unit _{3 -1} of the first pattern, the difference value calculating unit _{1 -1} by the difference value calculated based on the first pattern BE, de, fe, based on he, the first pattern By referring to the correction value table _2-1 , the correction value corresponding to the first pattern is obtained. In this case, the correction value acquisition unit _{3 -1} of the first pattern, four difference values BE, de, fe, four correction values Tbe corresponding to he, Tde, Tfe, obtains and outputs the The.

The correction value acquisition unit 3 _-2 of the second pattern, the difference value calculating unit 1 _-2 by the second based on the pattern calculated difference value bh, based on df, the correction value of the second pattern table By referring to _2-2 , a correction value corresponding to the second pattern is obtained. In this case, the correction value acquisition unit _{3 -2} of the second pattern, two difference values bh, two correction values Tbh corresponding to df, determines and outputs a Tdf.

The correction value acquisition unit _3-3 of the third pattern, based differential value calculating section _{1 -3} by a third, based on the pattern calculated difference value ae, ce, ge, the ie, the third pattern by referring to the correction value table 2 _-3, to obtain a correction value corresponding to the third pattern. In this case, the correction value acquisition unit _3-3 of the third pattern, the four difference values ae, ce, ge, four correction values Tae corresponding to ie, Tce, Tge, obtains and outputs the Tie.

  The pixel value conversion unit 4 performs a pixel value conversion process on the pixel value of the target pixel E set at the center in the processing region using the correction values obtained for each of the plurality of patterns by the correction value acquisition unit 3. Do. Various calculation contents of the pixel value conversion process can be applied. For example, pixel value conversion is performed by adding all of a plurality of correction values obtained by the correction value acquisition unit 3 to the pixel value of the target pixel E.

  In this case, it is also conceivable that the added value exceeds the range that can be taken as the pixel value. For example, if one pixel value is represented by 8-bit data, the possible range of the pixel value is 0 to 255. Therefore, when the added value becomes smaller than 0, the pixel value conversion unit 4 replaces the value with 0. When the added value becomes larger than 255, the value is replaced with 255.

  The pixel value conversion unit 4 outputs the pixel value obtained by the pixel value conversion process for the target pixel E as described above as the corrected pixel value of the target pixel E. By sequentially scanning the processing region over the entire range of the image and obtaining and outputting the corrected pixel value of each pixel of interest, it is possible to obtain an image subjected to pixel value conversion processing, that is, an image to which a desired image effect is given. it can.

  Hereinafter, the types of patterns used when applying various image effects and the contents of the correction value table will be described.

<Defocus Correction (Outline Enhancement)>
In the blur correction, an image having a sharp sense of depth is converted by increasing a luminance difference between ambiguous contour boundaries. Specifically, in an edge portion corresponding to a contour having a certain degree of luminance difference between images, pixel values for one to two lines from the boundary are added and emphasized (brighter) or darker in a lighter (brighter) area or darker ( By subtracting and emphasizing (making darker) the pixel values for one or two lines from the boundary in the darker area, the outline boundary is made to stand out and blurring is alleviated.

Here, the contents of the correction values set in the correction value tables 2 ₋₁ , 2 ₋₂ ,..., 2- _k for blur correction will be described in detail. As the correction value table _2-1 for the first pattern, table information having contents as shown in FIG. 5 is used. The table information is set such that the smaller the luminance difference value is, the larger the correction value is, and the larger the difference value is, the smaller the correction value is. The difference value and the correction value are linear functions (solid lines in the figure). ). Also, the correction value is positive when the difference value is negative, the correction value is negative when the difference value is positive, and the correction value is also zero when the difference value is zero. Is set.

  The linear function here does not require linearity in a strict sense. In the case of a linear function having linearity in a strict sense, it is necessary to have 511 correction values as table information corresponding to 511 difference values of −255 to 255. On the other hand, the difference values are grouped for each predetermined range (for example, −255 to −240, −239 to −224,..., 240 to 255), and a correction value is set for each group. Also good. The same applies to the linear functions described below.

  Note that each correction value of the table information indicated by the solid line graph in FIG. 5 is merely an example, and the present invention is not limited to this. That is, in FIG. 5, the hatched range surrounded by the alternate long and short dash line indicates a range that is preferably taken as the correction value. As long as it is within the range surrounded by the alternate long and short dash line, a correction value different from that indicated by the solid line graph in FIG. 5 can be used. For example, when it is desired to reduce the degree of emphasis, each correction value is set so that the inclination of the straight line is reduced. When it is desired to increase the degree of emphasis, each correction value is set so that the inclination of the straight line is increased. Further, it is not always necessary to take the form of a linear function, and a curve such as a quadratic function may be used. In addition, it is not always necessary to set the correction value to a value other than zero over the entire range where the difference value is −255 to 255 (excluding the point where the difference value is 0), for example, a range where the absolute value of the difference value is larger than a predetermined value Then, the correction value may be zero.

Also, as the correction value table _2-2 for the second pattern, table information having contents as shown in FIG. 6 is used. Here, FIG. 6A is used when the pixel values of the lighter (brighter) area and the darker (darker) area are subtracted and emphasized (darker) at the edge corresponding to the contour of the image. Table information. FIG. 6B is a table used when the pixel values of the lighter (brighter) area and the darker (darker) area are added and emphasized (brighter) at the edge corresponding to the contour of the image. Information.

  The table information shown in FIG. 6A is set so that the correction value increases as the luminance difference value decreases, and the correction value decreases as the difference value increases. The relationship is represented by a function (solid line in the figure). However, since the absolute value of the difference value is taken in the second pattern as described in FIG. 3, the difference value does not become a negative value, but only the correction value when the difference value becomes a positive value. Is set. In FIG. 6A, the correction value is set to a negative value when the difference value is a positive value, and the correction value is set to zero when the difference value is zero.

  The table information shown in FIG. 6B is set such that the smaller the difference value, the smaller the correction value, and the larger the difference value, the larger the correction value. The difference value and the correction value are linear functions ( The relationship is represented by a solid line in the figure. Again, only the correction value when the difference value becomes a positive value is set. In FIG. 6B, the correction value is set to a positive value when the difference value is a positive value, and the correction value is also set to zero when the difference value is zero. In FIG. 6, the hatched range surrounded by the alternate long and short dash line indicates a range that is preferably adopted as the correction value. As long as it is within the range surrounded by the alternate long and short dash line, a correction value different from that indicated by the solid line graph in FIG. 6 can be used.

As the correction value table 2 _-3 of the third pattern uses the table information of contents as shown in FIG. The table information is set such that the smaller the luminance difference value is, the larger the correction value is, and the larger the difference value is, the smaller the correction value is. The difference value and the correction value are linear functions (solid lines in the figure). ). Also, the correction value is positive when the difference value is negative, the correction value is negative when the difference value is positive, and the correction value is also zero when the difference value is zero. Is set. The third pattern correction value table 2 _-3 may be the same as the correction value table 2 _-1 of the first pattern. In FIG. 7, the hatched range surrounded by the alternate long and short dash line indicates a range that is preferably taken as a correction value. As long as it is within the range surrounded by the one-dot chain line, a correction value different from that shown in the solid line graph in FIG. 7 can be used.

Figure 8 is a diagram for explaining enhancement line by the correction value table 2 _-3 correction value table 2 _-1 and the third pattern of the first pattern. Here, FIG. 8A illustrates an emphasis line based on the correction value table _2-1 of the first pattern. In FIG. 8A, reference numeral 11 denotes a first processing area, and 12 denotes a second processing area obtained by shifting this by one pixel in the horizontal direction. Reference numeral 11E denotes a target pixel in the center of the first processing area 11, and reference numeral 12E denotes a target pixel in the center of the second processing area 12.

  As described with reference to FIG. 2, in the first pattern, the target pixels 11E and 12E set at the center in the processing regions 11 and 12, and the 4 positions positioned in the vertical and horizontal directions with respect to the target pixels 11E and 12E. Difference values from two neighboring pixels (adjacent pixels) are calculated. The arrow X1 indicates that, in the first processing region 11, a difference value is calculated between the target pixel 11E and the adjacent pixel above it. An arrow X2 indicates that, in the second processing region 12, a difference value is calculated between the target pixel 12E and the adjacent pixel above it.

In the first processing region 11, since the difference value between the pixel indicated by the arrow X1 is 10, the correction value corresponding to the value obtained from the correction value table 2 _-1 of the first pattern. On the other hand, since the difference value between the pixels indicated by the arrow X2 is zero, the correction value is also zero. As can be seen, the correction value table _2-1 of the first pattern is used for contour enhancement for one line.

On the other hand, FIG. 8 (b) illustrate the emphasis line based on the correction value table 2 _-3 of the third pattern. As described with reference to FIG. 4, in the third pattern, the target pixels 11E and 12E set at the center in the processing regions 11 and 12 and the four peripheral pixels positioned obliquely with respect to the target pixels 11E and 12E Difference values from (adjacent pixels) are calculated. An arrow Y1 indicates that, in the first processing region 11, a difference value is calculated between the target pixel 11E and an adjacent pixel diagonally above and to the left. An arrow Y2 indicates that, in the second processing region 12, a difference value is calculated between the target pixel 12E and an adjacent pixel diagonally above and to the left.

In the first processing region 11, since the difference value between the pixel indicated by arrow Y1 is 10, the correction value corresponding to the value obtained from the correction value table 2 _-3 of the third pattern. Similarly, since the difference value is also 10 between pixels indicated by the arrow Y2, the correction value corresponding to the value obtained from the correction value table 2 _-3 of the third pattern. As can be seen, the correction value table 2 _-3 of the third pattern is used for the edge enhancement of the two lines.

The emphasis lines in the correction value table _{2-2 for} the second pattern are not shown, but are used for contour emphasis for one line as in the correction value table _2-1 for the first pattern. In particular, the correction value table _2-2 of the second pattern is used for fine adjustment of the enhancement degree.

Figure 9 is a diagram showing an example of a different edge enhancement pattern using the correction value table 2 _-1 to 2 _-3 of each pattern described above in image manner. In FIG. 9, the horizontal direction indicates the pixel position, and the vertical direction indicates the pixel value. A thick solid line indicates a pixel value in each pixel of the original image, and a one-dot chain line indicates a pixel value in each pixel of the emphasized image after the pixel value conversion processing.

As indicated by the thick solid line, the pixel value of the original image has a relatively large difference in luminance value at a certain pixel position. This corresponds to the contour portion of the image. Contour enhancement by a predetermined calculation using the pixel values of two lines before and after, and a correction value stored in the correction value table 2 _-1 to 2 _-3 of each pattern such contour as a boundary I do. 21a, 21b show a line of edge enhancement based on the correction value table _{2 -1} of the first pattern, 22a, 22b are the two lines of edge enhancement based on the correction value table _{2 -3} of the third pattern Show. Note that, here, the outline enhancement for one line based on the correction value table _2-2 of the second pattern is omitted.

For example, in FIG. 9A, based on the correction value table _2-1 for the first pattern, the luminance difference for one line in contact with the boundary of the contour is in the plus direction (bright direction) and in the minus direction (dark direction). The contour is emphasized so that it becomes larger. Similarly, based on the correction value table 2 _-3 of the third pattern, the luminance difference between two lines of contact with the boundary of the contour is the contour emphasis so also increases in the negative direction to the positive direction.

  As a result, the outline emphasis of each of the first pattern and the third pattern is added for one line from the outline boundary, resulting in strong emphasis, and the second line is more than the outline emphasis based on the third pattern. Is weak (close to the pixel value of the original image). In this way, the image is sharpened by increasing the degree of emphasis of the contour portion, and a natural image in which the pixel value changes relatively smoothly from the pixel of the emphasized contour portion to its surrounding pixels. can do.

By changing the correction value of the correction value table 2 _-1 to 2 _-3, it is also possible to perform edge enhancement in a variety of patterns as illustrated in FIG. 9 (b) and FIG. 9 (c). Incidentally, FIG. 9B shows a pattern in which the contour is emphasized so that the luminance difference between the first and second lines in contact with the contour boundary increases only in the plus direction. FIG. 9C shows a pattern in which the emphasis directions are different between the first pattern and the third pattern. In this case, the correction value table 2 _-1 of the first pattern, and acts to produce a blurring effect, has weakened the first line of emphasis by the correction value table 2 _-3 of the third pattern.

As described above, in the present embodiment, the correction value tables 2 ₋₁ to 2 _-k having different patterns are used according to the positional relationship between the two pixels taking the difference, and both the positional relationship between the pixels and the difference value thereof are used. The correction value corresponding to the is obtained. Thereby, instead of simply determining the correction value according to the difference value, it is possible to select a pixel position and determine an appropriate correction value. Therefore, the image quality of the image obtained by the pixel value conversion process using the correction value can be improved as compared with the conventional technique in which the correction value is simply changed to change the enhancement degree.

Further, the width of the emphasis line at the contour boundary can be changed by changing the combination of the correction value tables 2 ₋₁ , 2 _{−2 ,.} For example, in terms of characters, the difference in luminance between the background and the character is always large, so the difference in pixel value (luminance value) between the pixel of interest in the center of the processing region and its surrounding pixels increases to some extent. For example, as shown in FIG. 10, in the region luminance difference is large between the target pixel and the surrounding pixels to emphasize the contour boundary at 2 line width by the correction value table 2 _-3 of the third pattern, when small luminance difference by highlighting the correction value table 2 _-1 by 1 line width of the first pattern, it can be easily viewed by the character more sharply.

  It is also possible to increase or decrease the emphasis width of the dark area by changing the emphasis ratio between the bright area and the dark area of the contour boundary. For example, in order to make black characters thick and easy to read, it is possible to give a reverse negative correction value without performing normal edge enhancement (plus correction) in a bright area. By using such a method, the image can be made more beautiful and the subtitles can be recognized more easily.

In the present embodiment, the degree of emphasis can be arbitrarily changed by changing the correction values of the correction value tables 2 ₋₁ , 2 _{−2 ,.} For example, if the brightness difference between the pixel of interest and the surrounding pixels is small, such as a thin freckle on the skin of a person, the enhancement level is weakened (the enhancement level is weakened in the range of difference values where correction is not desired). The correction value of each correction value table 2 ₋₁ , 2 ₋₂ ,..., 2- _k is determined), and other luminance differences are enhanced by increasing the degree of emphasis, thereby avoiding emphasis on freckles and emphasizing the outline A clear image is obtained. Further, since the mosquito noise, block noise, and the like generated by image compression (JPG, MPEG, etc.) do not have a large luminance difference, it is preferable to reduce the enhancement degree for such a small luminance difference.

  FIG. 11 is a diagram illustrating an example of an image to which an image effect of defocus correction is actually applied using the method of the present embodiment. FIG. 11A is an original image, and FIG. 11B is a defocus corrected image.

<Blur>
The blur image effect is to smooth the change in the pixel value at the boundary of the contour by correcting the difference in the pixel value of the edge corresponding to the contour of the image to be small. With this blurring effect, blurring of 1 to 4 lines at the boundary between the contours can be realized with one algorithm. This blur effect can also be used as a filter function. For example, the pixel of interest set at the center in the processing area protrudes compared to the surrounding pixels and is relaxed only by bright particles, only protrudes dark particles, relaxes both bright and dark particles, and is generated by image compression. Mosquito noise and block noise can be reduced.

Here, the details of the correction values set in the correction value tables 2 ₋₁ , 2 ₋₂ ,..., 2- _k for blurring will be described in detail. As the correction value table _2-1 for the first pattern, table information having contents as shown in FIG. 12 is used. The table information is set such that the smaller the luminance difference value is, the smaller the correction value is, and the larger the difference value is, the larger the correction value is. The difference value and the correction value are linear functions (solid lines in the figure). ). The correction value is negative when the difference value is negative, the correction value is positive when the difference value is positive, and the correction value is also zero when the difference value is zero. Is set. In FIG. 12, the hatched range surrounded by the alternate long and short dash line indicates a range that is preferably taken as a correction value. As long as it is within the range surrounded by the one-dot chain line, a correction value different from that indicated by the solid line graph in FIG. 12 may be used.

Further, as the correction value table _2-2 of the second pattern, table information having contents as shown in FIG. 13 is used. As described above, since the absolute value of the difference value is taken in the second pattern, the difference value never becomes a negative value, and only the correction value when the difference value becomes a positive value is set. Yes. In FIG. 13, the correction values are all set to zero regardless of the difference value. The fact that all correction values are zero means that the correction value table _2-2 of the second pattern is not used. However, if fine adjustment of the blurring effect is desired, a value other than zero may be set. In this case, a preferable correction value range is the same as that shown in FIG.

As the correction value table 2 _-3 of the third pattern uses the table information of contents as shown in FIG. 14. The table information is set such that the smaller the difference value, the smaller the correction value, and the larger the difference value, the larger the correction value. The difference value and the correction value are linear functions (solid lines in the figure). The relationship is expressed. The correction value is negative when the difference value is negative, the correction value is positive when the difference value is positive, and the correction value is also zero when the difference value is zero. Is set. The third pattern correction value table 2 _-3 may be the same as the correction value table 2 _-1 of the first pattern. In FIG. 14, the hatched range surrounded by the alternate long and short dash line indicates a range that is preferably taken as a correction value. As long as it is within the range surrounded by the one-dot chain line, a correction value different from that indicated by the solid line graph in FIG. 14 may be used.

More first sets the correction value as the pattern of the correction value table 2 _-1 and the third pattern blurring the line by the correction value table 2 _-3 may be considered in the same manner as in FIG. Correction value table 2 _-3 of the third pattern without, With only correction value table 2 _-1 of the first pattern, it is possible to impart a simple blurring effect of 1-2 lines in the edge boundary. You can also use the correction value table 2 _-3 of the third pattern in addition to the correction value table 2 _-1 of the first pattern, it is possible to impart blur effect over 3-4 line.

Further, as the contents of the table information used in the correction value table _2-1 of the first pattern, instead of the information shown in FIG. 12, the one shown in FIG. 15 can be used to give a filter effect to the image. it can. For example, it is possible to reduce mosquito noise and block noise generated by image compression. Specifically, the blurring degree is increased in the range where the difference value is around ± 1 to ± 32, and the blurring degree is slightly reduced in the other difference value ranges.

  FIG. 16 is a diagram illustrating an example of an image that is actually provided with the blurring effect using the method of the present embodiment. Here, FIG. 16A is an original image, and FIG. 16B is a blurred image.

<Blur + Bokeh correction>
The blur + defocus correction is to smooth an image that has been roughened by the blur correction, and to reduce the defocus feeling caused by the defocus correction. This image effect functions as a filter, and correction can be performed when the pixel of interest at the center is lighter (brighter) than peripheral pixels, when it is darker (darker), or both. In addition, mosquito noise generated by image compression, freckles with thin skin, small moles, etc. can be alleviated.

Here, the contents of the correction values set in the correction value tables 2 ₋₁ , 2 ₋₂ ,..., 2− _k for blur + defocus correction will be described in detail. As the correction value table _2-1 for the first pattern, table information having contents as shown in FIG. 12 is used (in this case, an unused example). Further, as the correction value table _2-2 of the second pattern, table information having contents as shown in FIG. 13 is used. As the correction value table 2 _-3 of the third pattern uses the table information of contents as shown in FIG.

In the case of functioning as a filter for reducing mosquito noise and freckles with thin skin, the correction value table _2-1 for the first pattern is shown in FIG. 15 instead of the table information having the contents as shown in FIG. Table information with contents as shown is used. Specifically, the difference value is ±. 1 to ± 32 before and after the range, the stronger the first correction value table 2 _-1 in the blurring degree of the pattern shown in Figure 7, the third pattern shown in FIG. 14 weakly set degree of correction value table 2 _-3 in out-of-focus correction. In other difference value ranges, the blur correction is strengthened and the blurring degree is slightly weakened. In this way, an easy-to-view sharp image can be obtained.

  FIG. 17 is a diagram illustrating an example of an image that is actually provided with blur + defocus correction using the method of the present embodiment. Here, FIG. 17A is an original image, and FIG. 17B is an image subjected to blur + blur correction.

<All color depth conversion>
In the all color density conversion, for example, a dark night image is brightened (lightened), or a bright white image is darkened (darkened) by converting the color of the entire image. By performing such density conversion, it is possible to make the image easier to see. Various color conversions can also be performed by making the RGB correction values unbalanced.

Here, the contents of the correction values set in the correction value tables 2 ₋₁ , 2 ₋₂ ,..., 2- _k for all color density conversion will be described in detail. The correction value table 2 _-3 correction value table 2 _-1 and the third pattern of the first pattern, using the table information of contents as shown in FIG. 18. In this table information, all correction values are set to the same value regardless of the difference value.

  In FIG. 18, the hatched range surrounded by the alternate long and short dash line indicates a range that is preferably taken as a correction value (all zeros are not included). As long as it is within the range surrounded by the one-dot chain line, a correction value different from that indicated by the solid line graph in FIG. 18 may be used. For example, it is not always necessary to set the correction value to the same value over the entire range of difference values. In this case, the correction value can be set to zero only in a partial area of the difference value.

Further, as the correction value table _2-2 for the second pattern, table information having contents as shown in FIG. 19 is used. In this table information, all correction values are set to the same value regardless of the difference value. However, only the correction value when the difference value becomes a positive value is set. In FIG. 19, the hatched range enclosed by the alternate long and short dash line indicates a range that is preferably taken as a correction value.

  FIG. 20 is a diagram illustrating an example of an image to which an image effect of all color depth conversion is actually applied using the method of the present embodiment. 20A is an original image, FIG. 20B is an image converted so that the density of the original image is high, and FIG. 20C is an image converted so that the density of the original image is low. is there.

<All color density conversion + defocus correction>
As described above, an example in which an image effect obtained by combining blur correction and defocus correction has been described. However, depending on the content of the correction value set in each correction value table 2 ₋₁ , 2 _{−2 ,.} Thus, it is possible to add an image effect in other combinations. One example is full color density conversion + out-of-focus correction.

For full color strength conversion + defocus correction, the correction value table 2 _-1 of the first pattern, using the table information of contents as shown in FIG. Further, as the correction value table _2-2 for the second pattern, table information having contents as shown in FIG. 19 is used. As the correction value table 2 _-3 of the third pattern uses the table information of contents as shown in FIG.

<All color density conversion + blur correction>
For full color strength conversion + blur correction, the correction value table 2 _-1 of the first pattern, using the table information of contents as shown in FIG. 12. Further, as the correction value table _2-2 for the second pattern, table information having contents as shown in FIG. 19 is used. As the correction value table 2 _-3 of the third pattern uses the table information of contents as shown in FIG. 14.

<All color density conversion + Blur + Defocus correction>
For full color strength conversion + Blur + defocus correction, the correction value table 2 _-1 of the first pattern, using the table information of contents as shown in FIG. 12. Further, as the correction value table _2-2 for the second pattern, table information having contents as shown in FIG. 19 is used. As the correction value table 2 _-3 of the third pattern uses the table information of contents as shown in FIG.

In addition, various variations other than the above can be achieved by changing the positional relationship between the two pixels for which the difference is taken and the contents of the correction values set in the correction value tables 2 ₋₁ , 2 _{−2 ,.} Rich image effects can be realized.

FIG. 21 is a diagram illustrating a fourth pattern as the positional relationship between two pixels that take a difference. In the fourth pattern, the difference value calculation unit 1 calculates the difference between the peripheral pixels B and H positioned in the vertical direction and the peripheral pixels D and F positioned in the horizontal direction of the target pixel E set at the center in the processing region. Calculate the value. That is, in the fourth pattern, the difference value calculation unit 1 calculates the following four difference values bd, bf, dh, and fh. Note that || indicates an absolute value.
bd = | B−D |, bf = | B−F |, dh = | D−H |, fh = | F−H |

FIG. 22 is a diagram illustrating a fifth pattern as the positional relationship between two pixels that take a difference. In the fifth pattern, the difference value calculation unit 1 calculates a difference value between pixels positioned in the relationship of shogi's Katsura jump among the pixels set in the processing region. That is, in the third pattern, the difference value calculation unit 1 calculates the following eight difference values af, ah, cd, ch, gb, gf, ib, and id.
af = A−F, ah = A−H, cd = C−D, ch = C−H
gb = GB, gf = GF, ib = IB, id = ID

<Negative / Positive contour extraction>
In negative / positive contour extraction, the range of absolute values of difference values is divided into two groups, and all correction values are set to the same value regardless of the difference values within the same group. Further, correction values belonging to different groups are set so that the difference between them is large. Hereinafter, the group with the smaller absolute value is referred to as the first group # 1, and the group with the larger absolute value is referred to as the second group # 2. If the difference value that is the transition between the first group # 1 and the second group # 2 is close to zero, a fine contour is extracted, and if it is far away, the fine contour is cut. Further, the width of the contour can be narrowed or thickened by a combination of the correction value tables 2 ₋₁ , 2 _{−2 ,.}

  FIG. 23 is a diagram showing an example of an image to which an image effect of negative / positive contour extraction is actually applied using the method of the present embodiment. Here, FIG. 23A is an original image, FIG. 23B is an image subjected to negative contour extraction, and FIG. 23C is an image subjected to positive contour extraction.

  Several style conversion effects described below can be realized by irregularly combining the above-described image effect providing functions. The point is that color density conversion correction values are set only for the first group # 1 with the smaller absolute value of the difference value using all color density conversion. This partial color density conversion does not emphasize the density of the entire image equally as in the case of full color density conversion, but performs irregular enhancement of the contour. In this case, the irregular emphasis is to give a strong correction value to generate a color change even though the difference between the pixel of interest and the surrounding pixels is small.

  The color change varies depending on the pattern of image data (RGB). For example, if the R color and G color are uniform data and the B color is irregular data, the R color and the G color are uniformly added and emphasized, and the B color is irregularly emphasized. As a result, color changes occur irregularly (in RGB). By adding out-of-focus correction, blur correction, etc. to this irregular emphasis, various style conversions are possible.

<Watercolor style conversion>
In the watercolor style conversion, the range of the absolute value of the difference value is divided into the first group # 1 having the smaller absolute value and the second group # 2 having the larger absolute value. Give a correction value. The watercolor-like image mentioned here is an image in which the outline is moderately black-enhanced without making the outline very sharp. Black emphasis can be realized by correction for blackening one to four lines of the contour boundary, and can be converted from a thin black contour line to a thick black contour line. It is also possible to add a stipple effect for the main image conversion. For example, the correction value may be set unbalanced in a predetermined region where the absolute value of the difference value between the correction value tables 2 ₋₁ , 2 _{−2 ,.}

  FIG. 24 is a diagram illustrating an example of an image to which an image effect of watercolor style conversion is actually applied using the method of the present embodiment. Here, FIG. 24A shows an original image, and FIG. 24B shows an image subjected to watercolor style conversion.

<Illustration-like conversion>
Even in illustration-like transformation, the absolute value range of the difference value is divided into the first group # 1 having the smaller absolute value and the second group # 2 having the larger absolute value, and is appropriate for each group. Correct correction value. The illustration-like image referred to here is one in which the color of the entire image is lighter than the watercolor painting and the outline is sharpened. Also in this image conversion, it is possible to add the above-described stippling effect.

  FIG. 25 is a diagram illustrating an example of an image to which an image effect of illustration-like conversion is actually applied using the method of the present embodiment. Here, FIG. 25A is an original image, and FIG. 25B is an image subjected to illustration-like conversion.

<Oil painting style conversion>
Also in oil painting style conversion, the range of absolute values of difference values is divided into a first group # 1 having a smaller absolute value and a second group # 2 having a larger absolute value, and appropriate for each group. Correct correction value. At this time, white outline enhancement is performed for the second group # 2. The oil painting-like image referred to here is obtained by blurring the outline, and performing defocus correction or white outline enhancement. White contour emphasis can be realized by correction for whitening one to four lines of the contour boundary, and can be converted from a thin white contour line to a thick white contour line. Also in this image conversion, it is possible to add the above-described stippling effect.

  FIG. 26 is a diagram illustrating an example of an image to which an image effect of oil painting style conversion is actually given using the method of the present embodiment. Here, FIG. 26 (a) is an original image, and FIG. 26 (b) is an image subjected to oil painting + stipple-style conversion.

<Fantasy transformation>
Even in fantasy-like transformation, the absolute value range of the difference value is divided into a first group # 1 having a smaller absolute value and a second group # 2 having a larger absolute value, and is appropriate for each group. Correct correction value. At this time, a correction value of −10 to −150 is given for the first group # 1. The illusion-like image referred to here is an image in which the overall image color of the oil painting style is made darker, contour blurring, defocus correction, and white contour enhancement are performed. Also in this image conversion, it is possible to add the above-described stippling effect.

  FIG. 27 is a diagram illustrating an example of an image to which an illusion-like conversion image effect is actually applied using the method of the present embodiment. Here, FIG. 27A is an original image, and FIG. 27B is an image that has been subjected to fantasy-like transformation.

As described above in detail, according to the present embodiment, the correction value tables 2 ₋₁ to 2 _-k having different patterns are used in accordance with the positional relationship between the two pixels taking the difference. A correction value corresponding to both the difference value is obtained. The contents of the correction values in each of the correction value tables 2 _{−1 to} 2 _-k can be arbitrarily set, and different correction values can be set for each table even if they are the same difference value. It is.

  In this way, it is possible to set an appropriate correction value according to the positional relationship between the pixels taking the difference, instead of simply determining the correction value according to the difference value. Specifically, even if the calculated difference values are the same, the correction values obtained can be made different depending on the positional relationship between the two pixels that have taken the difference. Also, depending on which pattern's positional relationship the difference value is obtained, that is, which pattern's table information is used to obtain the correction value, the obtained correction value can be made different and appropriate.

  Thereby, the image quality of the image obtained by the pixel value conversion process using the correction value can be improved. In addition, abundant variations of image effects that have never existed before can be provided by using table information or a combination thereof. Since an appropriate correction value can be determined by selecting the pixel position, an appropriate correction value is applied to the pixel corresponding to the interpolation position when the image is enlarged or reduced while applying various image effects as described above. As a result, jaggies can be prevented from occurring by smoothly interpolating between the emphasized pixels.

  The image processing apparatus according to the present embodiment described above can be realized by any of a hardware configuration, a gate array, an FPGA (Field Programmable Gate Array), and software. For example, when realized by software, the image processing apparatus according to the present embodiment is actually configured by including a CPU or MPU of a computer, RAM, ROM, and the like, and can be realized by operating a program stored in the RAM or ROM. .

  Therefore, it can be realized by recording a program that causes a computer to perform the functions of the above-described embodiment on a recording medium such as a CD-ROM and causing the computer to read the program. As a recording medium for recording the program, a flexible disk, a hard disk, a magnetic tape, an optical disk, a magneto-optical disk, a DVD, a nonvolatile memory card, and the like can be used in addition to the CD-ROM. It can also be realized by downloading the program to a computer via a network such as the Internet.

  Further, in order to realize the functions of the image processing apparatus according to the present embodiment in a network environment, all or a part of the programs may be executed by another computer. In addition, the functions of the above-described embodiments are realized by executing a program supplied by a computer, and the program is used in cooperation with an OS (operating system) or other application software running on the computer. When the functions of the above-described embodiment are realized, or when all or part of the processing of the supplied program is performed by a function expansion board or a function expansion unit of the computer, the function of the above-described embodiment is realized. Such a program is included in the embodiment of the present invention.

  In addition, the said embodiment is only what showed the example of actualization in implementing this invention, and, as a result, the technical scope of this invention should not be interpreted limitedly. In other words, the present invention can be implemented in various forms without departing from the spirit or main features thereof.

  The present invention is useful for an apparatus that imparts an image effect by performing predetermined pixel value conversion processing on image data.

It is a functional block diagram which shows the structural example of the image processing apparatus by this embodiment. It is a figure which shows a 1st pattern as a positional relationship of two pixels which take a difference. It is a figure which shows a 2nd pattern as a positional relationship of two pixels which take a difference. It is a figure which shows a 3rd pattern as a positional relationship of two pixels which take a difference. It is a figure which shows the example of the content of the correction value table of the 1st pattern used for a blurring correction | amendment. It is a figure which shows the example of the content of the correction value table of the 2nd pattern used for a blurring correction | amendment. It is a figure which shows the example of the content of the correction value table of the 3rd pattern used for a blurring correction | amendment. It is a figure for demonstrating the emphasis line by the correction value table of a 1st and 3rd pattern. It is a figure which shows the example of various outline emphasis patterns using the correction value table of each pattern. It is a figure which shows the example of a combination of the correction value table of a 1st pattern and a 3rd pattern. It is a figure which shows the example of the image which actually gave the image effect of a blurring correction | amendment using the method of this embodiment. It is a figure which shows the example of the content of the correction value table of the 1st pattern used for blurring correction | amendment. It is a figure which shows the example of the content of the correction value table of the 2nd pattern used for blurring correction | amendment. It is a figure which shows the example of the content of the correction value table of the 3rd pattern used for blurring correction. It is a figure which shows the example of the content of the correction value table of the 1st pattern used for filters. It is a figure which shows the example of the image which provided the image effect of the blurring correction | amendment actually using the method of this embodiment. It is a figure which shows the example of the image which provided the image effect of the blur + defocus correction | amendment actually using the method of this embodiment. It is a figure which shows the example of the content of the correction value table of the 1st and 3rd pattern used for all color density conversion. It is a figure which shows the example of the content of the correction value table of the 2nd pattern used as an object for all color density conversion. It is a figure which shows the example of the image to which the image effect of all the color depth conversion was actually provided using the method of this embodiment. It is a figure which shows a 4th pattern as a positional relationship of two pixels which take a difference. It is a figure which shows a 5th pattern as a positional relationship of two pixels which take a difference. It is a figure which shows the example of the image which provided the image effect of negative / positive outline extraction actually using the method of this embodiment. It is a figure which shows the example of the image which actually gave the image effect of watercolor style conversion using the method of this embodiment. It is a figure which shows the example of the image which provided the image effect of illustration-like conversion actually using the method of this embodiment. It is a figure which shows the example of the image which provided the image effect of oil painting style conversion actually using the method of this embodiment. It is a figure which shows the example of the image which actually gave the image effect of the fantasy style conversion using the method of this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Difference value calculating part 2 Table information storage part 3 Correction value acquisition part 4 Pixel value conversion part

Claims (5)

In a processing area set for m pixels × n pixels (m and n are each an integer of 1 or more and at least one is 2 or more) constituting a part of the input image, a predetermined positional relationship is established. A difference value calculation unit for calculating a difference value between two pixel values,
A table information storage unit storing correction values corresponding to the difference values as table information;
A correction value acquisition unit that refers to the table information based on the difference value calculated by the difference value calculation unit and obtains a correction value according to the difference value;
A pixel value conversion unit that performs a pixel value conversion process on the pixel value of the pixel of interest in the processing region using the correction value obtained by the correction value acquisition unit;
The difference value calculation unit is configured to prepare a plurality of patterns as the predetermined positional relationship and calculate the difference value for each pattern.
The table information storage unit stores a plurality of table information storing different correction values for the plurality of patterns,
The correction value acquisition unit refers to the table information related to the same pattern as the pattern for which the difference value is obtained based on the difference value calculated for each pattern by the difference value calculation unit, and performs correction according to the difference value. An image processing apparatus characterized in that a value is obtained. The difference value calculation unit is, as one of the plurality of patterns, a difference value between a target pixel set at the center in the processing region and peripheral pixels positioned in the vertical and horizontal directions with respect to the target pixel. The image processing apparatus according to claim 1, wherein: The difference value calculation unit sandwiches a difference value between two neighboring pixels positioned in the vertical direction across the target pixel set in the center of the processing region and the target pixel as one of the plurality of patterns. The image processing apparatus according to claim 1, wherein a difference value between two neighboring pixels positioned in the horizontal direction is calculated. The difference value calculation unit calculates, as one of the plurality of patterns, a difference value between a target pixel set at the center in the processing region and a peripheral pixel positioned in an oblique direction with respect to the target pixel. The image processing apparatus according to claim 1. In a processing area set for m pixels × n pixels (m and n are each an integer of 1 or more and at least one is 2 or more) constituting a part of the input image, a predetermined positional relationship is established. A difference value calculation means for calculating a difference value of pixel values for a certain two pixels, and calculating the difference value for each of a plurality of patterns prepared as the predetermined positional relationship;
A table information storage unit that stores a correction value corresponding to the difference value as a plurality of table information for each of the plurality of patterns based on the difference value calculated for each of the plurality of patterns by the difference value calculating means. Using the correction value obtaining means for obtaining a correction value according to the difference value from the table information relating to the same pattern as the pattern for which the difference value was obtained, and the correction value obtained by the correction value obtaining means, Pixel value conversion means for performing pixel value conversion processing on the pixel value of the target pixel in the region;
An image processing program for causing a computer to function as each means.