JP4192671B2 - Image processing apparatus and image processing method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and an image processing method for synthesizing an input image on another image in accordance with a mask pattern for designating effective pixels at the time of synthesis from the input image, and in particular, performing an enlargement / reduction process on the input image. The present invention relates to an image processing apparatus and an image processing method for synthesizing on another image.
[0002]
[Prior art]
In recent years, images are handled as digital data, and images generated by computer graphics, received television images, playback images from recorded data such as optical disks, etc. are enlarged and reduced, and these images are combined with other images to create new ones. An image processing apparatus that generates a simple image is generally used. For example, some television receivers have a function of enlarging / reducing menu screens and character strings stored in advance in an internal memory, and combining and displaying them on a received television image. Also, a pixel having a specific value for luminance, color data, etc. is taken out from the enlarged / reduced image, and is subjected to special processing such as replacement of color data or transparency to be combined with another image. ing.
[0003]
When enlarging an image, interpolation of pixels according to the enlargement ratio is performed, and when reducing, reduction of pixels according to the reduction ratio is performed. In this case, for example, if the image is enlarged by arranging the same pixels as the original image or the image is reduced by simply thinning out the pixels, the image quality of the generated image often deteriorates significantly. For this reason, after performing a filter process that performs a calculation using a predetermined coefficient based on these pixel values using pixels in a predetermined section of the original image corresponding to the enlarged and reduced pixels, another image is obtained. Is generally synthesized.
[0004]
As a general conventional example for correcting pixel data of a generated image based on pixel data in a predetermined section in the original image, there is a hierarchical data generation device as follows. This hierarchical data generation device is a device that performs hierarchical encoding to generate image signals of a plurality of layers having different resolutions, and generates pixels in the upper layer based on the average value of the pixel values of adjacent four pixels in the lower layer. By replacing one pixel of the adjacent four pixels in the lower layer with the generated upper layer pixel, errors generated when the upper layer is generated are dispersed and superimposed in the lower layer data, and the amount of data is reduced. Compressed (see, for example, Patent Document 1).
[0005]
[Patent Document 1]
Japanese Unexamined Patent Publication No. 2000-92331 (paragraph numbers [0049] to [0061], FIG. 5)
[0006]
[Problems to be solved by the invention]
However, when an image subjected to the above-described filter processing at the time of enlargement / reduction is combined with another image, the color of the boundary portion of the image blurs and becomes an blurred image, and the image quality may deteriorate. Hereinafter, an image generated using such filter processing will be specifically described. Here, an image when an image is enlarged will be described as an example.
[0007]
FIG. 8 is a diagram illustrating an example of a conventional image when an image is enlarged and combined with another image.
FIG. 8 shows a case where the image 101 on which the text “A” of the alphabet is drawn is enlarged like the image 102 and is synthesized with the background image 103 to generate the synthesized image 104. When combining images, a mask pattern is used to indicate whether or not each pixel in the original image is to be displayed on the combined image. For example, the mask pattern value (mask value) when displaying is “0”, the mask value when not displaying is “1”, and the mask value of the display area of the text “A” in the image 101 of FIG. 1 ”, and the mask value of the other region is“ 0 ”. By using such a mask pattern, a composition process is performed such that the background image 103 is displayed in the display area of the text “A”.
[0008]
By the way, when the image 101 is enlarged, it is necessary to generate a mask pattern corresponding to the enlargement ratio. In the image 102 obtained by enlarging the image 101, the mask value at the boundary portion of the display area of the text “A” becomes a value between “0” and “1” by the filtering process. In FIG. 8, interpolation is performed based on two adjacent pixels in the image 101, and the mask value at the boundary is calculated as “0.8” by performing an operation using a coefficient for the pixel value. . Thereby, in the synthesized image 104, the image 102 and the background image 103 are synthesized at a ratio of 2: 8 at the boundary of the display area of the text “A”. Therefore, at this boundary portion, the color of the background image 103 and the color of the original text “A” are mixed and blurred, and a clear display is not obtained.
[0009]
A similar phenomenon can occur when images are reduced and combined. In addition, in an image composed of thin lines and points such as font data, a phenomenon of image quality deterioration such as disappearance of lines and points occurs when the same filter processing is performed.
[0010]
The present invention has been made in view of such problems, and an object of the present invention is to provide an image processing apparatus capable of preventing deterioration in image quality when an image is enlarged and reduced and combined with another image. To do.
[0011]
Another object of the present invention is to provide an image processing method capable of preventing deterioration of image quality when an image is enlarged and reduced and combined with another image.
[0012]
[Means for Solving the Problems]
  In the present invention, in order to solve the above-described problem, in an image processing apparatus that synthesizes the input image on another image in accordance with a mask pattern for designating effective pixels at the time of synthesis from the input image, the input image and the mask pattern The first and second enlargement / reduction means for executing enlargement / reduction processing accompanied by filter processing, and the respective mask values of the mask pattern output from the second enlargement / reduction means for each pixel A mask pattern conversion means for determining whether or not the pixel is the effective pixel and converting the pixel into a new mask pattern.Then, the mask pattern conversion means compares the mask value of each pixel with a predetermined threshold value and determines whether or not to make the effective pixel based on the comparison result. It is arbitrarily set according to the enlargement / reduction ratio in the first and second enlargement / reduction means.An image processing apparatus is provided.
[0013]
  Here, the first and second enlargement / reduction means execute enlargement / reduction processing with filter processing on the input image and the mask pattern, respectively. The mask pattern conversion means determines whether or not to make each of these pixels effective pixels according to the mask value of each pixel of the mask pattern output from the second enlargement / reduction means, and converts it to a new mask pattern. To do. Therefore, the converted new mask pattern is composed of binary mask values converted according to the mask value of each pixel, and the color at the boundary portion of each image is combined with other images. Will not mix.Further, the mask pattern conversion means compares the mask value of each pixel of the mask pattern from the second enlargement / reduction means with a predetermined threshold value, and determines whether or not to make it an effective pixel based on the comparison result. To do. At this time, the threshold value used for the comparison is arbitrarily set according to the enlargement / reduction ratio in the first and second enlargement / reduction means.
[0014]
  Further, according to the present invention, image processing for synthesizing the input image on another image according to a mask pattern for designating effective pixels at the time of synthesis from the input imageapparatus, An enlargement / reduction process with a filter process is performed on each of the input image and the mask pattern.Output from the first and second enlargement / reduction means and the second enlargement / reduction meansAccording to the mask value of each pixel of the mask patternBeforeDetermine whether each pixel is the effective pixel,Convert to new mask patternAnd a mask pattern conversion unit that compares the mask value of each pixel with a predetermined threshold value and determines whether or not to determine the effective pixel based on the comparison result. The threshold value is arbitrarily set according to the input image.Image processing characterized byapparatusIs provided.
[0015]
  Here, the first and second enlargement / reduction means execute enlargement / reduction processing with filter processing on the input image and the mask pattern, respectively. The mask pattern conversion means determines whether or not to make each of these pixels effective pixels according to the mask value of each pixel of the mask pattern output from the second enlargement / reduction means, and converts it to a new mask pattern. To do. Therefore, the converted new mask pattern is composed of binary mask values converted according to the mask value of each pixel, and the color at the boundary portion of each image is combined with other images. Will not mix. Further, the mask pattern conversion means compares the mask value of each pixel of the mask pattern from the second enlargement / reduction means with a predetermined threshold value, and determines whether or not to make it an effective pixel based on the comparison result. To do. At this time, the threshold used for comparison is arbitrarily set according to the input image.The
[0016]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram illustrating a configuration example of an image processing apparatus according to an embodiment of the present invention.
[0017]
The image processing apparatus shown in FIG. 1 is an apparatus for generating a composite image obtained by scaling an input image and combining it on a background image. As an input image, for example, an image generated by cutting out a predetermined area from an image stored in a memory, a television image received by a television tuner, an image read out from a recording medium such as an optical disk, or the like is applied. It is possible.
[0018]
The image processing apparatus includes a key determination unit 11 that determines whether each pixel of an input image matches a key, a mask pattern generation unit 12 that generates a mask pattern according to a determination result of the key determination unit 11, Enlargement / reduction units 13 and 14 for performing enlargement / reduction processing with filter processing on the input image and the mask pattern, respectively, and a mask pattern for re-determining effective pixels from the enlarged / reduced mask pattern and converting them to a new mask pattern The conversion unit 15, the output pixel selector 16 that outputs the effective pixels of the input image according to the converted mask pattern, and the output pixels are synthesized on the background image according to the mask pattern, and the synthesized image data (hereinafter, synthesized image) (Referred to as data).
[0019]
The key determination unit 11 receives image data of an input image (hereinafter referred to as input image data) and a threshold value of a pixel value for extracting a specific pixel from the input image. This threshold value is called a key, and is input as a parameter indicating an upper limit value and a lower limit value of a specific pixel value. The key determination unit 11 determines that this pixel matches the key when each pixel value of the input image falls within the range from the upper limit value specified by the key to the lower limit value, and determines the determination result for each pixel as a mask pattern generation unit. 12 is output.
[0020]
The mask pattern generation unit 12 generates a mask pattern for extracting pixels to be combined from the input image based on the determination result of the key determination unit 11. The mask value of each pixel of the mask pattern generated here is constituted by binary data indicating whether each pixel is valid or invalid as an extraction target. The mask pattern generation unit 12 can be set with a key mode for designating whether a pixel that matches the key is valid as an extraction target or a pixel that does not match is valid.
[0021]
The enlargement / reduction unit 13 receives data of each pixel of the input image and performs enlargement / reduction processing of the input image according to the designation of the enlargement / reduction ratio. At the time of enlargement, pixel interpolation corresponding to the enlargement ratio is performed, and at the time of reduction, pixels are thinned out according to the reduction ratio. During such enlargement / reduction processing, by using a plurality of adjacent pixels in the original input image corresponding to each pixel after enlargement / reduction, an operation using a predetermined coefficient is performed based on these pixel values. A filter process for determining the pixel value after enlargement / reduction is performed.
[0022]
The enlargement / reduction unit 14 performs an enlargement / reduction process on the mask pattern output from the mask pattern generation unit 12 at the same enlargement / reduction ratio as the enlargement / reduction unit 13. This enlargement / reduction process is basically the same as the process executed on the input image data in the enlargement / reduction unit 13. The mask value of the mask pattern output from the enlargement / reduction unit 14 takes multivalued data by the filter process executed during the enlargement / reduction process.
[0023]
The mask pattern conversion unit 15 determines again whether or not each pixel is valid according to the mask determination criterion information from the mask value of the mask pattern output from the enlargement / reduction unit 14, and sets the mask value as an effective pixel and an invalid pixel. Are converted into binary data indicating the above and output as a new mask pattern. The mask determination criterion information is information for setting a valid / invalid determination criterion for the mask value of the input mask pattern.
[0024]
The output pixel selector 16 receives the input of the image data from the enlargement / reduction unit 13 and the mask pattern from the mask pattern conversion unit 15, and outputs the pixel data validated by the mask pattern from the enlarged / reduced image. . As a result, an image obtained by extracting a predetermined area from the image obtained by enlarging or reducing the input image is generated. Further, according to the setting of the pixel replacement value and the pixel replacement mode, the pixel value of the valid pixel can be replaced with a predetermined value and output. For example, the replacement pixel value is information that specifies a pixel value for changing the color of the effective pixel to a predetermined color, and the pixel replacement mode is information that specifies whether or not to replace the specified pixel value. is there.
[0025]
The composition processing unit 17 receives the image data from the output pixel selector 16, the mask pattern from the mask pattern conversion unit 15, and the image data of the background image (hereinafter referred to as background image data), and receives the output pixel selector. The image data output from 16 is combined with the background image data according to the mask pattern, and the combined image data is output. As a result, a composite image in which a predetermined area of the image extracted by the output pixel selector 16 is composite-displayed on the background image is generated.
[0026]
In this image processing apparatus, an enlargement / reduction process with a filter process is performed on an input image, and a similar enlargement / reduction process is also performed on a mask pattern generated from the input image. Then, the mask pattern that has been subjected to the enlargement / reduction process is converted into a new mask pattern by the mask pattern conversion unit 15, and the input image that has been subjected to the enlargement / reduction process is synthesized on the background image using the converted mask pattern. Here, the mask value of the mask pattern that has been subjected to the enlargement / reduction processing is converted into binary data in accordance with the mask determination reference information in the mask pattern conversion unit 15, thereby preventing deterioration in image quality at the boundary portion of each image to be synthesized. This is the most characteristic point of the present invention.
[0027]
Next, the operation of the image processing apparatus will be specifically described.
In the present embodiment, as an example, it is assumed that the pixel value of the input image is represented by a luminance component and a color difference component (blue, red). In the following description, it is assumed that each component of the pixel value is Yin, Uin, Vin, and these are each represented by 8-bit data. The key input to the key determination unit 11 is similarly designated as 8-bit data for the upper limit value and the lower limit value of each component. Hereinafter, for each component of the pixel value, the upper limit is expressed as Ymax, Umax, Vmax, and the lower limit is expressed as Ymin, Umin, Vmin.
[0028]
FIG. 2 is a diagram showing bit allocation for each component of the pixel value for the input key.
As shown in FIG. 2, the keys indicating the upper limit value and the lower limit value of the pixel value are each input to the key determination unit 11 as 24-bit data. For the upper limit of the pixel value, Ymax is assigned to the upper 8 bits, Umax is assigned to the next 8 bits, and Vmax is assigned to the next 8 bits. Similarly, the lower limit value is assigned in the order of Ymin, Umin, and Vmin in increments of 8 bits from the top.
[0029]
The key determination unit 11 determines whether or not the pixel value of each pixel of the input image data matches a condition specified by the key. Specifically, when Yin, Uin, and Vin satisfy all of the following expressions (1) to (3), it is determined that the condition is met.
[0030]
[Expression 1]
Ymax ≧ Yin ≧ Ymin (1)
Umax ≧ Uin ≧ Umin (2)
Vmax ≧ Vin ≧ Vmin (3)
Whether or not the pixel value of the input image matches the specified key as described above is determined for each pixel, and the determination result is output to the mask pattern generation unit 12. The mask pattern generation unit 12 sets a mask value for each pixel according to the determination result of the key determination unit 11, and generates a mask pattern. The mask value is a value indicating whether each pixel of the input image is valid or invalid as a pixel to be extracted. Here, “0” is set when valid, and “1” is set when invalid. Shall be. This means that the pixel with the mask value “0” is synthesized on the background image and the pixel with the mask value “1” is designated not to be synthesized.
[0031]
Further, the mask pattern generation unit 12 designates whether the pixel determined to match the key by the key determination unit 11 is valid or whether the pixel that does not match is valid according to the key mode. Here, the former is called a key match mode, and the latter is called a key knot match mode.
[0032]
FIG. 3 is a diagram showing the correspondence between the determination result of the key determination unit and the mask signal for each key mode.
In the case of the key-not-match mode, as shown in FIG. 3A, the mask signal level is set to “1” indicating invalidity for the pixel that matches the key by the determination in the key determination unit 11, and the matching is performed. The level of the mask signal is output as “0” indicating that it is valid for the pixels that are not. On the other hand, in the key match mode, as shown in FIG. 3B, the mask signal level is set to “0” for the pixel that matches the key, and the mask signal is set for the pixel that does not match. The level is output as “1”. Thus, by specifying the key mode, the extraction area specified by the key can be inverted.
[0033]
The mask pattern output from the mask pattern generation unit 12 is input to the enlargement / reduction unit 14. The enlargement / reduction unit 14 performs an enlargement / reduction process with a filter process on the input mask pattern in accordance with the designated enlargement / reduction ratio. For example, when the image is enlarged by interpolating the pixels, or when the image is reduced by thinning out the pixels, the filtering process is performed by a plurality of adjacent pixels corresponding to the enlarged and reduced pixels among the pixels of the original mask pattern. This is a process of calculating the mask value of the pixel after the enlargement / reduction by performing an operation using a predetermined coefficient based on the mask value.
[0034]
Here, FIG. 4 is a diagram for explaining an example of calculating a mask value when enlarging an image.
FIG. 4 shows an example in which the mask value after the enlargement process is calculated from the mask values of two adjacent pixels in the original mask pattern during the enlargement process. In addition, it is assumed that the enlargement ratio is set so that the section for five pixels becomes the section for about seven pixels. In this figure, for example, a section from pixel a to pixel b of the original mask pattern is interpolated by pixel c in the enlarged mask pattern. In this case, the mask value of the pixel c is calculated by performing an operation using a coefficient based on the mask values of two adjacent pixels on the original mask pattern corresponding to the pixel c, that is, the pixel a and the pixel b. The If the mask values of the pixels a, b, and c are P (a), P (b), and P (c), respectively, and the coefficients applied to the pixels a and b are coefa and coefb, respectively, the mask value P ( c) is calculated by the following equation (4).
[0035]
[Expression 2]
P (c) = P (a) * coefa + P (b) * coefb (4)
The enlargement / reduction unit 14 sequentially shifts the pixels of the mask pattern after the enlargement process, repeatedly performs the calculation using the above equation (4), and sequentially calculates the mask value. Also, such calculation is performed in both the horizontal direction and the vertical direction. FIG. 4 exemplifies calculation results when coefa and coefb are set to “0.3” and “0.7”, respectively. Note that FIG. 4 shows a case where the above calculation is performed only on the pixels in the horizontal direction for the sake of simplicity.
[0036]
By performing such a filter process, as shown in FIG. 4, the pixel after the enlargement process corresponds to a section in which two pixels having mask values “0” and “1” in the original mask pattern are adjacent to each other. In this case, the mask value of the pixel after the enlargement process takes an intermediate value between “0” and “1”. Such a pixel is an area where the original image and the background image are mixed and displayed in the composite image generated using the enlarged mask pattern. In FIG. 4, at the time of synthesis, the pixels on which the enlarged input image is displayed are shown in white, the pixels on which the background image is displayed are shown in black, and the input image and the background image are mixed and displayed. Pixels are indicated by diagonal lines.
[0037]
On the other hand, the enlargement / reduction unit 13 calculates the pixel value of the image after the enlargement / reduction processing by performing the same calculation as that of the enlargement / reduction unit 14 on each component of the pixel value of the input image data. Thereby, the positional relationship between each pixel of the image output from the enlargement / reduction unit 13 and each pixel of the mask pattern output from the enlargement / reduction unit 14 is 1: 1. Therefore, in this state, in the image after the enlargement / reduction processing, the pixel corresponding to the pixel whose mask value takes an intermediate value between “0” and “1” is displayed in the background image at the composition ratio according to the mask value at the time of composition. The pixel is specified to be mixed with the corresponding pixel, and this pixel causes image quality deterioration at the time of composition.
[0038]
Therefore, the mask pattern that has been subjected to the enlargement / reduction processing is converted by the mask pattern conversion unit 15 so that the mask value indicates either valid or invalid based on the mask determination reference information, whereby the original image and the background image are converted. Eliminates mixed areas and prevents image quality degradation. In this embodiment, as an example, it is assumed that two determination modes are designated by mask determination reference information. One is a determination mode in which only pixels with a mask value of “0” are valid and all others are invalid, and the other is in which only pixels with a mask value of “1” are invalidated. It is a judgment mode that makes all other than valid
FIG. 5 is a diagram illustrating an example of mask values converted according to each determination mode.
[0039]
FIG. 5 shows changes in the mask value according to the determination mode for pixels in the same region as that shown in FIG. The mask value of the mask pattern enlarged by the enlargement / reduction unit 14 takes a value from “0” to “1” as shown in FIG. On the other hand, in the determination mode in which only the pixel whose mask value is “0” is valid, an intermediate mask value between “0” and “1” is designated as shown in FIG. The mask value of the pixel that has been changed is converted to “1”. Further, in the determination mode in which only the pixel having the mask value “1” is invalid, as shown in FIG. 5C, the pixel for which an intermediate mask value between “0” and “1” has been designated. The mask value is converted to “0”.
[0040]
The output pixel selector 16 selectively outputs each pixel of the image output from the enlargement / reduction unit 13 according to the mask pattern converted as described above. Here, when the pixel replacement mode is set so as not to perform the pixel value replacement, whether the signal of each pixel is output 100% or not at all according to the mask value of the mask pattern. In any case, when synthesized with the background image by the synthesis processing unit 17, there is no pixel displayed by mixing the color of the background image and the color of the original image. Therefore, in the synthesized image, the boundary portion between the synthesized images is clearly displayed, and the image quality is improved.
[0041]
By the way, the output pixel selector 16 can replace the pixel value of the validated pixel with a predetermined value and output it. When the pixel replacement mode is set to perform pixel replacement, for each pixel with a mask value of “0”, each component of the pixel value is replaced with the value specified by the replacement pixel value and output. The In the mask pattern input to the output pixel selector 16, all mask values take either “0” or “1”, so that the color of the replaced pixel region is clearly displayed and mixed with the background image. There is nothing.
[0042]
Next, the transition of the image and the mask pattern will be described using a specific image example with reference to FIGS. In FIGS. 6 and 7, it is assumed that an image in which the alphabet “A” is drawn as the text is enlarged, the display area of the text “A” is cut out, and the background image appears in the area. To do.
[0043]
FIG. 6 is a diagram showing an example of mask pattern transition in this case.
First, the key determination unit 11 designates a pixel value corresponding to the color of the text “A” as a key, and the mask pattern generation unit 12 sets a key knot match mode to match the key. A mask pattern 61 is obtained in which the area that is not to be designated is designated as an effective pixel. In this mask pattern 61, that is, the mask value of the pixels constituting the text “A” is “1” indicating invalidity, and the mask values of the other pixels are “0” indicating validity.
[0044]
Such a mask pattern 61 is enlarged by the enlargement / reduction unit 14. In the mask pattern 62 after processing, the mask value is an intermediate value between “0” and “1” for the pixel at the boundary portion of the display area of the text “A” in accordance with the calculation by the filter processing. In FIG. 6, as an example, the mask value of this pixel is set to “0.8”.
[0045]
Next, the mask pattern 62 is input to the mask pattern conversion unit 15. Here, for example, when a judgment mode in which only a pixel having a mask value of “0” is valid and all others are invalid is set by the mask judgment reference information, an intermediate mask between “0” and “1”. All the values are converted to “1”, and a mask pattern 63 including only valid pixels and invalid pixels is generated.
[0046]
FIG. 7 is a diagram illustrating an example of image transition corresponding to FIG.
As described above, the text “A” is drawn in the input image 71. The input image 71 is enlarged by the enlargement / reduction unit 13 to generate an image 72. In the image 72, the color of the text “A” and the color of the adjacent area are mixed in the boundary portion of the display area of the text “A” with the calculation by the filter processing. Then, the output pixel selector 16 selectively outputs each pixel according to the converted mask pattern 63 with respect to such an image 72.
[0047]
As shown in FIG. 7, the mask pattern 63 is designated as an invalid pixel including an area of the text “A” and a portion where the colors of the area and the adjacent area are mixed in the enlarged image 72. Therefore, in the synthesized image 74, the color of the region other than the text “A” of the input image 71 and the color of the background image 73 are not mixed in the boundary region, and the boundary of each part is clearly displayed. .
[0048]
As described above, in the above-described image processing apparatus, the mask value of the mask pattern that has been subjected to the enlargement / reduction process with the filter process is converted into binary data in accordance with the mask determination criterion information by the mask pattern conversion unit 15 and output. In the pixel selector, output selection of each pixel of the image that has been enlarged / reduced according to the converted mask pattern is performed. As a result, at the boundary portion between the enlarged / reduced image and the background image, the colors of the images are not mixed and the image is not unclear, and a high-quality composite image can be generated.
[0049]
Further, by replacing the pixel value of the effective pixel in the output pixel selector 16, it is possible to replace the color of the area extracted from the input image, or to apply a special effect such as performing a transparency process on this area. However, even in such a case, a high-quality composite image with a clear boundary with the background image can be generated.
[0050]
In addition, the enlargement / reduction processing function used in the above-described image processing apparatus has been generally used in the past, and by using such a general enlargement / reduction processing function, a high-performance by a relatively simple configuration. Since a high-quality composite image is generated, an effect of reducing component costs can be obtained.
[0051]
By the way, in the above-described embodiment, the mask pattern conversion unit 15 makes two determinations, when only a pixel whose mask value is “0” is enabled and when only a pixel whose mask value is “1” is disabled. Mode was used. In these judgment modes, for example, when text is extracted from an input image with a key, whether a mask pattern is generated in the key match mode or generated in the key knot match mode based on this extraction region Can be used by switching.
[0052]
In general, when the text itself is synthesized on the background image by the key match mode, the impression of visual recognition is better when the text is thinly displayed. Therefore, it is preferable to specify a determination mode in which only a pixel having a mask value of “0” is valid to narrow the effective pixel area. On the other hand, when the background image is displayed by cutting out the text display area by the key-knot match mode, the impression at the time of visual recognition is better when the area is enlarged. Therefore, it is preferable to specify a determination mode in which only pixels having a mask value of “1” are invalid to widen the effective pixel area.
[0053]
In addition to these two determination modes, for example, an intermediate value between “0” and “1” is specified as a threshold value of a mask value for determining whether each pixel is valid or invalid. May be. Further, this threshold value may be specified according to the state of the input image data, the enlargement / reduction ratio, and the like.
[0054]
The larger the threshold value, the larger the number of effective pixels and the wider the area. Conversely, the smaller the threshold value, the smaller the number of effective pixels and the smaller the area. For example, when the key match mode is set for the mask pattern generation unit 12, when the extraction region from the input image is configured by thin lines or small points, or even if the lines are thick, For example, when the distance is short, if the number of effective pixels is increased, fine lines and points may be crushed. Therefore, in such a case, it is generally preferable to reduce the threshold value.
[0055]
Further, when such an input image is enlarged, the threshold value can be increased because lines and points are not easily collapsed. However, considering the case of reduction, when there are many fine lines and points, it is necessary to increase the display area by increasing the threshold value so that they do not disappear. On the other hand, when the line is thick to some extent and the distance from the adjacent line is short, the line may be connected due to the reduction, so the threshold may need to be reduced.
[0056]
As an example, when extracting text from an input image with a key, a method of setting the above threshold value for each font of the text can be considered. Furthermore, the threshold value may be changed up and down by a predetermined amount in accordance with the enlargement / reduction ratio for each font or the selection of the key mode.
[0057]
When the input image data is input as an individual file, the above threshold value may be set according to the selection of the file. Further, the ratio of pixels in the input image of the extracted region based on the key may be calculated, and the threshold value may be automatically set according to this ratio.
[0058]
As described above, the image processing apparatus can easily extract various images from the input image according to the combination of the key designation and the key mode designation, and is optimal for the extracted image. Since a simple threshold value can be set in the mask pattern conversion unit 15, it can be suitably used for a wide range of applications.
[0059]
【The invention's effect】
As described above, according to the image processing apparatus of the present invention, the new mask pattern converted by the mask pattern conversion means is constituted by the binary mask value converted according to the mask value of each pixel. As a result, when the image is combined with another image, colors are not mixed at the boundary portion of each image and unclear, so that the image quality of the combined image can be improved.
[0060]
  Also,The mask pattern conversion means compares the mask value of each pixel of the mask pattern from the second enlargement / reduction means with a predetermined threshold value and determines whether or not to make it an effective pixel based on the comparison result, The threshold used at this time is arbitrarily set according to the enlargement / reduction ratio in the first and second enlargement / reduction means, or according to the input image, so that the image quality of the synthesized image can be increased. One layerCan be improved.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration example of an image processing apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram showing bit allocation for each component of a pixel value for an input key.
FIG. 3 is a diagram illustrating a correspondence between a determination result of a key determination unit and a mask signal for each key mode.
FIG. 4 is a diagram illustrating an example of a method for calculating a mask value when enlarging an image.
FIG. 5 is a diagram illustrating an example of mask values converted according to each determination mode;
FIG. 6 is a diagram illustrating an example of transition of a mask pattern.
FIG. 7 is a diagram illustrating an example of image transition.
FIG. 8 is a diagram illustrating an example of a conventional image when an image is enlarged and combined with another image.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 11 ... Key determination part, 12 ... Mask pattern production | generation part, 13, 14 ... Enlargement / reduction part, 15 ... Mask pattern conversion part, 16 ... Output pixel selector, 17 ... Composition processing part

Claims (4)

In an image processing apparatus for synthesizing the input image on another image according to a mask pattern for designating effective pixels at the time of synthesis from the input image,
First and second enlargement / reduction means for executing enlargement / reduction processing with filter processing on each of the input image and the mask pattern;
Mask pattern conversion means for determining whether or not each pixel is to be the effective pixel according to a mask value of each pixel of the mask pattern output from the second enlargement / reduction means, and converting the pixel into a new mask pattern; ,
I have a,
The mask pattern conversion means compares the mask value of each pixel with a predetermined threshold value and determines whether or not to make the effective pixel based on the comparison result. the image processing apparatus according to claim Rukoto be optionally set depending on the scaling rate in the first and second scaling means. In an image processing apparatus for synthesizing the input image on another image according to a mask pattern for designating effective pixels at the time of synthesis from the input image,
First and second enlargement / reduction means for executing enlargement / reduction processing with filter processing on each of the input image and the mask pattern;
Mask pattern conversion means for determining whether or not each pixel is to be the effective pixel according to a mask value of each pixel of the mask pattern output from the second enlargement / reduction means, and converting the pixel into a new mask pattern; ,
Have
The mask pattern conversion means compares the mask value of each pixel with a predetermined threshold value and determines whether or not to make the effective pixel based on the comparison result. An image processing apparatus that is arbitrarily set according to an image. In an image processing method for compositing the input image on another image according to a mask pattern for designating effective pixels at the time of compositing from the input image,
  For each of the input image and the mask pattern, an enlargement / reduction process with a filter process is performed,
  According to the mask value of each pixel of the mask pattern on which the enlargement / reduction processing has been performed, it is determined whether or not each pixel is the effective pixel, and is converted into a new mask pattern,
  Using the converted new mask pattern and the image data of the input image on which the enlargement / reduction processing has been executed, a synthesis process with the other image is performed,
  In the process of converting to the new mask pattern, the mask value of each pixel is compared with a predetermined threshold value, and it is determined whether or not to be the effective pixel based on the comparison result. Is arbitrarily set according to the enlargement / reduction ratio in the enlargement / reduction process. In an image processing method for compositing the input image on another image according to a mask pattern for designating effective pixels at the time of compositing from the input image,
  For each of the input image and the mask pattern, an enlargement / reduction process with a filter process is performed,
  According to the mask value of each pixel of the mask pattern on which the enlargement / reduction processing has been performed, it is determined whether or not each pixel is the effective pixel, and is converted into a new mask pattern,
  Using the converted new mask pattern and the image data of the input image on which the enlargement / reduction processing has been executed, a synthesis process with the other image is performed,
  In the process of converting to the new mask pattern, the mask value of each pixel is compared with a predetermined threshold value, and it is determined whether or not to be the effective pixel based on the comparison result. Is arbitrarily set according to the input image.

Images