JP6606660B2 - Image data encoding device - Google Patents

Description

  The present invention relates to an image data encoding apparatus that compresses and encodes image data that is synthesized by superimposing an image in which transparency information is set with another image.

  Conventionally, various methods for compressing and encoding still image data and moving image data are known (see Non-Patent Document 1 below).

  For example, the video compression standard H.264. 264, H. In a standard codec such as H.265, there is a function (adaptive quantization) that can individually specify a quantization coefficient for each image region in a compressed stream. By specifying a quantization coefficient for each image area in the compressed stream, it is possible to increase the data compression rate while maintaining data that reproduces a fine image.

  By adaptive quantization, a region where noise is conspicuous in an image is quantized with low compression, and a region where noise is not conspicuous is quantized with high compression, thereby making it possible to achieve both high image quality and high compression as a whole.

  As a method for selecting a quantization coefficient, the encoder automatically measures the complexity by edge detection or brightness detection in the image, and determines the quantization coefficient for each region.

  In Patent Document 1, a quantization coefficient is selected according to each component of YUV in color system image data including a luminance component and a color difference component. By selecting a quantization coefficient table according to YUV components, data compression can be performed at a compression rate suitable for the ratio of each component.

JP 2009-206566 A

Toshiba Review Vol. 64 no. 6 (2009) p7-10

  However, there are many image frames synthesized regardless of still images and moving images by utilizing an alpha blend method in which a plurality of images are superimposed using an α value indicating transparency information. Videos that use alpha blending are used in gaming machines such as pachinko machines, for example, and images that display multiple numbers and characters representing the lottery results, images that display characters, etc. are superimposed on the background image. Used when composing an image frame.

  Even in an image using such an alpha blend, each image is compressed by a standard codec such as JPEG or MPEG. Therefore, even if the image data is not displayed at all by alpha blending, it is a certain amount of compressed data. As a result, the image data tends to be large, and the conventional compression method has room for improvement.

  The present invention has been made in view of the above problems, and it is an object of the present invention to provide an image encoding apparatus capable of improving the compression rate while suppressing deterioration in appearance in an image that is superimposed and synthesized. It is said.

In order to solve this problem, the invention described in claim 1 compresses and encodes color information data in which transparency information is set for each pixel and transparency data as the transparency information data. An image data encoding apparatus comprising: a color information encoder that encodes color information data; and a transparency data encoder that encodes transparency data. A predetermined calculation is performed on the predetermined quantization coefficient from the initial coefficient setting unit with respect to the transparency reference correction number according to the transparency of the pixel to be processed and the pattern reference correction number according to the pattern state of the matrix including the pixel. A coefficient control unit for calculating a quantization coefficient and determining a divisor to be given to the quantization unit, wherein the coefficient control unit is configured to reduce the opacity of a pixel set based on the transparency information by half. The divisor is double With the settings to perform, transmittance of the pixel is characterized by being configured so as to set to the permeability reference correction number to a constant value during the predetermined value or more.

Invention according to claim 2, in addition to the structure according to claim 1, used in the predetermined operation in the coefficient control unit, the size of the range of values of the permeability reference correction numbers, the picture reference By setting the range of the numerical value of the correction number to be wider than the range of the numerical value of the correction number, the size of the correction range of the predetermined quantization coefficient by the transparency reference correction number is the predetermined quantization coefficient by the pattern reference correction number. The correction range is set to be larger than the size of the correction range .

  According to the first aspect of the present invention, when the frequency-converted image data is compressed and encoded, the image data is compressed in accordance with the transparency. The compression rate of the image and the compression rate of the image that is easily visible can be adjusted according to the transmittance. Therefore, it is possible to provide an image data encoding apparatus capable of improving the compression rate while suppressing deterioration in appearance in an image that is superimposed and synthesized.

  According to the second aspect of the present invention, since the higher the transparency of the image for which the transparency information is set, the higher the compression rate of the image data of the image, the more highly transparent the image is difficult to see. The image data of the image in which the transparency information is set can be greatly compressed while suppressing the appearance deterioration in the processed image.

  In addition, the lower the transparency of an image for which transparency information is set, the higher the compression rate of the image data of the other image, the more difficult it is to see the other image hidden behind one of the images with low transparency. The image data of other images can be greatly compressed while suppressing the deterioration of the appearance.

  Therefore, the higher the transparency of an image with transparency information set, the lower the compression rate of the image data of other images, or the higher the compression rate of the image data of the image with transparency information set. By doing so, it is possible to improve the compression rate without causing appearance deterioration in the superimposed moving image.

According to the first and second aspects of the present invention, the color information encoder includes: a color information encoder that encodes color information data; and a transparency data encoder that encodes transparency data. , A predetermined calculation with respect to a predetermined quantization coefficient from the initial coefficient setting unit, a transparency reference correction number according to the transparency of the corresponding pixel and a pattern reference correction number according to the pattern state of the matrix including the pixel A coefficient control unit that calculates a quantization coefficient by performing and determining a divisor to be given to the quantization unit, and the coefficient control unit is configured to reduce the opacity of the pixel set based on the transparency information by half. The divisor is set to be doubled, and when the transparency of the pixel is equal to or greater than a predetermined value, the transparency reference correction number is set to be a constant value. Quantization coefficient depending on degree It can be set. Therefore, it is easy to compress the image data at a compression rate suitable for the transparency.

It is the schematic explaining the structure in the encoding apparatus which concerns on embodiment of this invention. It is a block diagram which shows the Y encoder, U encoder, and B encoder in the encoding apparatus which concerns on embodiment of this invention. It is a block diagram which shows the coefficient setting part in the encoding apparatus which concerns on embodiment of this invention. It is a block diagram which shows the (alpha) encoder in the encoding apparatus which concerns on embodiment of this invention. It is a table | surface which shows the size standard after the compression with respect to the quantization coefficient of the encoding apparatus which concerns on embodiment of this invention. It is a table | surface which shows the transparency reference | standard correction number with respect to ( alpha) value of the encoding apparatus which concerns on embodiment of this invention. It is a table | surface which shows the correction number with respect to the abundance of the high frequency component of the encoding apparatus which concerns on embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. Although the present embodiment will be described based on an example applied to an MPEG encoding method, the present invention may be applied to any image encoding method other than MPEG.

  FIG. 1 shows the configuration of the entire encoder in the encoding apparatus according to the embodiment, FIGS. 2 and 4 are block diagrams showing the encoder of the encoding apparatus, and FIG. 3 is a block diagram showing the coefficient setting unit of the encoding apparatus. It is.

  An image to be compressed by the encoding apparatus according to the present embodiment is an image in which each frame is configured by superimposing and synthesizing one image for which transparency information is set and the other image.

  Each image is composed of image data having color information and transparency information for each pixel.

  The color information may be, for example, RGB data, but in the present embodiment, YUV data including a luminance component and a color difference component is used.

  The information on the transparency is a value that represents a numerical value in multiple stages from a completely opaque state to a completely transparent state, and is set for each pixel. In the present embodiment, α data obtained by converting an α value indicating a transparency of 100% to 0% is set.

  An encoding apparatus 10 as an “image data encoding apparatus” shown in FIG. 1 is an apparatus that compresses and encodes such image data in accordance with transparency. In the encoding device 10, image data is processed in units of a matrix composed of a predetermined number of pixels.

  In a moving image in which one image and the other image are superimposed, when compressing image data according to the transparency, even if the image data of one image is compressed according to the transparency of one image, the other The image data of the images may be compressed, or the image data of both images may be compressed.

  In the present embodiment, an example of an apparatus that compresses image data of one image according to the transparency of the one image will be described.

  As shown in FIG. 1, the encoding device 10 of the present embodiment includes a plurality of encoders 11, and an input image input from the input side is compressed and encoded by the plurality of encoders 11, so that the output side It is configured to be output as a bit stream.

  The plurality of encoders 11 include a Y encoder 11y that receives Y data and α data and compresses and encodes Y data, and a U encoder 11u that receives U data and α data and compresses and encodes U data. A V encoder 11v that receives V data and α data and compresses and encodes the V data; and an α encoder 11α that receives α data and compresses and encodes the α data.

  As shown in FIG. 2, the Y encoder 11y, the U encoder 11u, and the V encoder 11v excluding the α encoder 11α are a DCT unit 12 that generates DCT data from image data, and a code generator that generates encoded data from DCT data. Unit 13, a coefficient setting unit 14 that sets a quantization coefficient from image data and transmits the quantization coefficient to code generation unit 13, a prediction processing unit 15 that performs inter-frame prediction processing using the quantization data of code generation unit 13, It is equipped with.

  The DCT unit 12 decomposes each image data into frequency components by performing frequency conversion such as discrete cosine transform in units of matrix, and generates DCT data that is coefficientized.

In the encoding device 10 of the present embodiment, since the prediction processing unit 15 is provided, the image data from the input side is input to the DCT unit 12 as a prediction error formed by a difference from the prediction image, and the prediction error DCT data is generated.

The code generation unit 13 generates and encodes quantized data by compressing image data processed by a predetermined encoding method based on the quantization coefficient. Here, the “predetermined encoding scheme” refers to a transform encoding scheme based on frequency conversion such as discrete cosine transform in the DCT unit 12 which is the immediately preceding configuration.

  The code generation unit 13 includes a quantization unit 16 that compresses DCT data to generate quantized data, and a variable length encoding unit 17 that further compresses and encodes the quantized data.

  The quantization unit 16 generates quantized data by compressing the DCT data transmitted from the DCT unit 12 after the frequency conversion of the image data based on the quantization coefficient set in the coefficient setting unit 14 to generate data. Reduce the amount. The quantization unit 16 can compress the DCT data according to the transparency by using the quantization coefficient set according to the transparency information by the coefficient setting unit 14 as described later.

  In the present embodiment, quantized data is generated by dividing and compressing DCT data obtained by frequency-converting each color information of YUV by a divisor set by the coefficient setting unit 14 corresponding to the quantization coefficient. .

  On the other hand, the variable length encoding unit 17 further compresses the quantized data by entropy encoding using a Huffman code, an arithmetic code, or the like to generate encoded data.

  The coefficient setting unit 14 sets the quantization coefficient by correcting the predetermined quantization coefficient of the initial coefficient setting unit 25 based on the image data, sets the divisor corresponding to the quantization coefficient, and sets the quantum of the code generation unit 13. To the conversion unit 16.

  Here, the divisor based on the quantized coefficient is a positive integer that increases as the quantized coefficient increases, and is a numerical value that can be compressed more appropriately than the quantized coefficient by dividing and rounding the DCT data.

  As illustrated in FIG. 3, the coefficient setting unit 14 includes a memory unit 21 that stores input image data, a pixel information determination unit 22 that generates a correction value from color information, and an α that generates a correction value from transparency information. A value determination unit 23; and a coefficient control unit 24 that corrects a predetermined quantization coefficient of the initial coefficient setting unit 25 and sets a quantization coefficient.

  The memory unit 21 stores image data transmitted from the input side.

  The pixel information determination unit 22 determines the state of the pattern in the matrix of the image data by appropriately using the peripheral image data for the image data of the matrix unit, and sets the correction value of the quantization coefficient according to the state of the pattern To do.

  In this determination by the pixel information determination unit 22, whether the state of the pattern of the matrix is a flat region with a small amount of high frequency components, an edge region with a large amount of high frequency components, or a smooth change region in between. , Etc. are determined in multiple stages.

  The pixel information determination unit 22 sets a correction value based on the determination result and transmits the correction value to the coefficient control unit 24.

  The pixel information determination unit 22 is configured such that correction values are preset in multiple stages so as to correspond to the state of the pattern of the matrix, and the correction values corresponding to the determination results can be selected. In the present embodiment, the design reference correction number for correcting the quantization coefficient by adding to the predetermined quantization coefficient set in the initial coefficient setting unit 25 is set.

  The pattern reference correction number is set to be larger in the flat area, smaller in the edge area, and intermediate in the smooth change area.

  The α value determination unit 23 determines α data in the transparency information of the input image data, and sets a quantization coefficient correction value according to the transparency.

  The α value determination unit 23 is configured such that correction values are set in advance in multiple stages so as to correspond to the transparency, and the correction value corresponding to the determination result can be selected. The correction value is set as a value for increasing the quantization coefficient so that the higher the image transparency, the higher the compression rate of the image data.

  The correction range of the correction value corresponding to the transparency is set wider than the correction range of the correction value corresponding to the state of the matrix pattern.

  In the present embodiment, the transparency reference correction number for correcting the quantization coefficient by adding to the default quantization coefficient set in the initial coefficient setting unit 25 is set. In the present embodiment, the transparency reference correction number is set so as to increase as the transparency increases.

  Further, the transparency reference correction number is set so that the quantization coefficient when the transparency in the image constituting the moving image frame is the smallest is set to a desired value.

  Further, the transmittance reference correction number is set so that the divisor is doubled when the transmittance is doubled and the transmittance is halved.

  In addition, by adding the transparency reference correction number to the predetermined quantization coefficient, the process may be simplified by setting the transparency reference correction number to a constant value for a range exceeding the range of the predetermined quantization coefficient.

  The coefficient control unit 24 sets a quantization coefficient, and transmits a divisor based on the quantization coefficient to the quantization unit 16 of the code generation unit 13 and an inverse quantization unit 28 of the prediction processing unit 15 described later.

  In the coefficient control unit 24, the default quantization coefficient set in the initial coefficient setting unit 25 is transmitted, and the default quantization coefficient is determined by the correction value from the pixel information determination unit 22 and the correction value from the α value determination unit 23. The quantization coefficient is set by correcting.

  In the coefficient control unit 24 according to the present embodiment, the α value determination unit 23 sets the pattern reference correction number set by the pixel information determination unit 22 according to the state of the matrix pattern and the transparency information of the image data. The predetermined quantization coefficient is corrected according to the transmitted transparency reference correction number.

  Specifically, the quantization coefficient is set by adding the pattern reference correction number and the transparency reference correction number to the predetermined quantization coefficient.

  The coefficient control unit 24 sets a divisor for dividing and compressing the DCT data based on the set quantization coefficient, and sets the divisor to the inverse quantization of the quantization unit 16 of the code generation unit 13 and the prediction processing unit 15. To the control unit 28.

  Such a divisor is used to generate quantized data by compressing DCT data by the quantizing unit 16 of the code generating unit 13 as described above.

  The prediction processing unit 15 decodes the quantized data and performs interframe prediction processing.

  The prediction processing unit 15 has a function of motion prediction and motion compensation. As shown in FIG. 2, the prediction processing unit 15 performs an inverse quantization unit 28 that performs an inverse quantization process on the quantized data, and an inverse of the decoded DCT data. A predicted image is generated from the inverse DCT unit 29 that performs DCT processing, the frame memory unit 31 that stores the image data of the previous frame including the decoded image data, and the image data from the input side and the image data of the previous frame A prediction processing execution unit 32.

  The inverse quantization unit 28 receives from the coefficient setting unit 14 a quantization coefficient obtained by correcting the predetermined quantization coefficient according to the state of the matrix pattern and the transparency of the image data. Using this quantization coefficient, the quantized data transmitted from the quantizing unit 16 is inversely quantized to decode the DCT data.

  The inverse DCT unit 29 decodes the image data from the frequency component by performing inverse DCT processing on the decoded DCT data.

  The frame memory unit 31 stores the image of the previous frame generated using the decoded image data, and provides the image to the prediction processing execution unit 32 as a reference image.

  The prediction process execution unit 32 generates a prediction image from the image data of the reference image from the frame memory unit 31 and the image data from the input side.

  The prediction image generated by the prediction processing execution unit 32 generates a prediction error based on the difference from the image data from the input side, and is compressed and encoded by the DCT unit 12 and the code generation unit 13. Further, by matching the image data decoded from the quantized data of the prediction error with the predicted image, an image of the previous frame is generated and used for the inter-frame prediction process.

  On the other hand, as shown in FIG. 4, the α encoder 11α includes a DCT unit 12 that generates DCT data from image data, a code generation unit 13 that generates encoded data from DCT data, and quantized data of the code generation unit 13. And a prediction processing unit 15 that performs inter-frame prediction processing using the.

  The α encoder 11α is not provided with the coefficient setting unit 14 like the Y encoder 11y, the U encoder 11u, and the V encoder 11v shown in FIG. Instead, the default quantization coefficient of the initial coefficient setting unit 25 common to the initial coefficient setting unit 25 of FIG. 3 in the Y encoder 11y, the U encoder 11u, and the V encoder 11v is the quantization unit 16 or the prediction processing unit of the code generation unit 13. It is configured to be transmitted to 15 inverse quantization units 28.

  Other configurations of the α encoder 11 are the same as those of the Y encoder 11y, the U encoder 11u, and the V encoder 11v.

  In such an α encoder 11α, it is possible to generate transparency compressed data based on α data of transparency information set for each pixel of image data.

  Transparency compressed data is used when decoding quantized data generated by compressing image data including color information and transparency information for each pixel, or encoded data obtained by further compressing quantized data. Is done.

  Next, the operation of such an encoding apparatus 10 will be described.

  As shown in FIG. 1, when image data of an input image is input to the encoding device 10, Y data of color information and α data of transparency information are converted into Y encoder 11 y, U encoder 11 u, In addition to being transmitted to the V encoder 11v, the α data of the transparency information is transmitted to the α encoder 11α, and the image data is encoded.

  As shown in FIG. 2, in each encoder 11 y, 11 u, 11 v, the image data of the input image is transmitted to the prediction processing unit 15, the coefficient setting unit 14, and the DCT unit 12. It is transmitted to the DCT unit 12 in the form of a prediction error.

  When the image data of the input image is transmitted to the prediction processing unit 15, an inter-frame prediction process is executed from the image data from the input side and the image data of the previous frame to generate a predicted image.

  In the coefficient setting unit 14 shown in FIG. 3, the quantization coefficient shown in FIG. 5 and the divisor set based on each quantization coefficient are set in advance.

  For example, assuming that the encoded data amount is about 4% of the image data of the input image, the default quantization coefficient is set to 12 as an initial setting of the initial coefficient setting unit 25 (FIG. 5). (See line 100).

  When the image data of the input image is transmitted to the coefficient setting unit 14, the pixel information determination unit 22 detects the frequency component included in each matrix and determines the state of the pattern in each matrix. Then, as shown in FIG. 6, a pattern reference correction number is set in accordance with the state of the matrix pattern.

  For example, when it is determined that the image has an extremely high frequency component and is a pattern in the edge region, -2 is set as the pattern reference correction number (see the row denoted by reference numeral 200 in FIG. 6) and is given to the coefficient control unit 24.

  In addition, the α value determination unit 23 shown in FIG. 3 determines α data of transparency information from the image data input to the coefficient setting unit 14. Then, as shown in FIG. 7, a transparency reference correction number corresponding to the transparency is set.

  In FIG. 7, α data is set in a range of 1 to 256 so as to correspond to a range of 100% to 0% of an α value indicating the transmittance. The α data has an inverse relationship with the α value. For example, α data is 1 when the α value is 100% at the maximum and the transmittance is highest, and α data is 256 when the α value is 0% and the transmittance is the lowest.

  For example, when the alpha data of the transparency information in the input image data is 64 or more and less than 128, +12 is set as the transparency reference correction number and is given to the coefficient control unit 24.

  In the coefficient control unit 24 shown in FIG. 3, the default quantization coefficient +12 of the initial setting value from the initial coefficient setting unit 25, the pattern reference correction number −2 from the pixel information determination unit 22, and the transmission from the α value determination unit 23. The degree reference correction number +12 is summed to set a divisor +22, which is given to the quantization unit 16 and the prediction processing unit 15 of the code generation unit 13.

  When the number of corrections is added to the predetermined quantization coefficient and exceeds the range of the predetermined quantization coefficient, the correction number is fixed to a constant value. For example, in FIG. 7, the transmittance reference correction number is fixed to +24 within a range where α data is 8 or less. That is, since a range of 0 to 32 is assumed as the range of the default quantization coefficient, even if a value of +24 or more is set as the transparency reference correction number when the α data is 8 or less, the default value is set as the default value. By combining with the quantization coefficient +12, the expected range of the predetermined quantization coefficient is exceeded. In this range, since the level of the encoded data set after the variable length encoding does not change much, the transparency reference correction number is fixed to +24 as in the region indicated by reference numeral 300 in FIG.

  On the other hand, as shown in FIG. 1, when image data of an input image is transmitted and transmitted to the DCT unit 12 as a prediction error consisting of a difference from the prediction image of the prediction processing unit 15, the prediction error is subjected to frequency conversion and DCT. Data is generated.

  When the DCT data is transmitted to the code generation unit 13, the quantization unit 16 compresses the DCT data using the divisor given from the coefficient setting unit 14 to generate the quantized data. In this quantization, image data is compressed for each quantization coefficient, for example, to the extent shown in the size guideline in FIG.

  Further, the variable length encoding unit 17 compresses and encodes the quantized data by entropy encoding, and outputs the encoded data as a bit stream. In this variable length coding, image data is compressed for each quantized coefficient, for example, to the extent shown in the size guideline in FIG.

  Then, such encoded data is output from the encoding device 10 as a bit stream.

  According to the encoding apparatus 10 of the present embodiment as described above, when image data is compressed and encoded, the image data is compressed according to the transparency. Specifically, the higher the transparency of one image for which transparency information is set, the greater the transparency reference correction number. As a result, the coefficient given from the coefficient control unit 24 to the quantization unit 16 is larger ( = The divisor is large), the compression rate of the image data of the image is high. As a result, it is possible to improve the compression rate while suppressing the appearance deterioration in the superimposed and synthesized image.

  In an image synthesized by superimposing one image with transparency set on the other image, one image with high transparency is displayed lighter than the other image, or the transparency is low. The other image is hidden behind one of the set images. That is, an image that is difficult to be viewed by a viewer is generated in a part of the superimposed images. Even if the image data of such an image is compressed and roughened, even if it is deleted if it is noticeable, the appearance of the superimposed and synthesized image does not deteriorate.

  In the present embodiment, the image data of one image is compressed according to the transparency of one image. Therefore, the higher the transparency of one image, the higher the compression rate of the image data of one image or the other image. The compression rate of the image data can be lowered.

  Therefore, when the transparency of one image is high and it becomes difficult to see, the image data of one image can be significantly compressed. On the other hand, when the transparency of one image is low and the image becomes easy to see, the image data deterioration of one image can be made clear by reducing the compression rate of the image data of one image.

  In the conventional adaptive quantization technology, the quantization coefficient is selected according to the RGB component or the quantization coefficient is selected according to the YUV component, and the transmittance is not referred to. For this reason, when the image is overlapped with the background image by alpha blending, even if the region is almost invisible, the encoding is performed so as to maintain the same image quality as the other regions.

  On the other hand, in the present embodiment, the higher the transparency of one image for which transparency information is set, the higher the compression rate of the image data of one image, or the compression rate of the image data of the other image. By reducing it, waste of compression efficiency can be greatly improved.

  In the moving picture coding apparatus 10 according to the present embodiment, the coefficient setting unit 14 is configured to correct a predetermined quantization coefficient set in advance according to the transparency of image data and set the quantization coefficient. Specifically, the quantization coefficient is set by adding the transparency reference correction number set in accordance with the transparency of the image data to the predetermined quantization coefficient.

  Therefore, the quantization coefficient can be increased or decreased according to the transparency, the image data can be easily compressed at a compression rate suitable for the transparency, and the structure for setting the quantization coefficient can be simplified.

  In addition, since the coefficient reference correction number is set so that the quantization coefficient when the transparency in the image is the smallest is set to a desired value in the coefficient setting unit 14, the compression can be performed within an appropriate range.

  Furthermore, when the α data representing the transparency is half the maximum value (256 in the present embodiment) (128 in the present embodiment), the divisor is doubled compared to the divisor of the α data maximum value, and Since the coefficient setting unit 14 sets the transparency reference correction number so that the opacity is doubled, that is, the divisor is doubled as the α data becomes 1/2 of the predetermined α data, the influence of the transmission is reduced. The compression rate of the received image data can be greatly adjusted.

  The moving picture coding apparatus 10 according to the present embodiment decodes the quantized data generated from the prediction error based on the quantized coefficient corrected according to the transparency and performs an inter-frame prediction process. It has.

  Therefore, even when image data is more highly compressed using transparency, inter-frame prediction processing can be performed appropriately, and without causing visual deterioration in the image of the video frame superimposed and synthesized, The compression rate can be improved.

  In the encoding device 10 of the present embodiment, the code generation unit 13 generates quantized data by compressing image data including color information and transparency information for each pixel, and compresses transparency information for each pixel. Thus, the transparency compressed data is generated.

  Therefore, by using the transparency compressed data, the encoded data encoded based on the quantized data can be appropriately decoded, and the appearance deterioration of the image in compression and decoding can be further suppressed.

[Modification]
Next, a modification of this embodiment will be described.

  In this modification, for each moving image frame synthesized by superimposing one image on which transparency information is set and the other image, the image data of the other image is converted according to the transparency of one image. It is an example of the encoding apparatus 10 of the moving image to compress. That is, when the transparency of the foreground image data is completely opaque in the superimposed image, the background portion where the image data is superimposed is not displayed at all, so the compression rate of the image data of the non-displayed background portion is increased. Setting the compression increases the data compression rate.

  The encoding device 10 in the modification is the same as that in the above embodiment except that the α value determination unit 23 of the coefficient setting unit 14 is different.

  In the α value determination unit 23 of the coefficient setting unit 14, the transparency of the image data for which the transparency information is set is input to the coefficient control unit 24 of the coefficient setting unit 14 that processes the image data of the other image. It is configured to be.

  The α value determination unit 23 determines the transparency of the image data of one image, and sets a correction value according to the transparency, specifically, a transparency reference correction number. At this time, the correction value is set as a value for decreasing the quantization coefficient so that the higher the transparency of one image, the lower the compression rate of the image data.

  The transparency reference correction number is transmitted to the coefficient control unit 24 of the coefficient setting unit 14 that processes the image data of the other image.

  In the coefficient control unit 24 of the other image, the transparency reference correction number set according to the transparency in the image data of the one image and the state of the matrix pattern in the other image transmitted from the pixel information determination unit 22 The quantization coefficient is set according to the pattern reference correction number set according to.

  Here, the quantization coefficient of the other image is set by adding the transparency reference correction number and the pattern reference correction number to the predetermined quantization coefficient from the initial coefficient setting unit 25, and the divisor is set based on this quantization coefficient. The

  The divisor set in this way is transmitted to the quantization unit 16, so that the DCT data generated by frequency conversion of the prediction error of the image data of the other image is divided by the divisor in the quantization unit 16. Then, the quantized data is generated and compressed, and further compressed by the variable length encoding unit 17, whereby the image data of the other image can be compressed and encoded.

  Even with the encoding device 10 according to the modified example as described above, the same effects as those of the above-described embodiment can be obtained.

  In addition, in this modification, the image data of the other image is compressed according to the transparency of the one image for which the transparency information is set. Therefore, the higher the transparency of one image, the more the image data of the other image. The compression rate can be lowered, or the compression rate of the image data of the other image can be increased as the transparency of one image is lower.

  Therefore, in the other image, the image data of an area that is hidden by one image and is difficult to see because the transparency of one image is low can be significantly compressed. On the other hand, image quality deterioration is suppressed by lowering the compression rate of the image data in the area where it is easy to see through one image because of the high transparency of the other image. can do.

  Moreover, it is also possible to use the said embodiment and the said modification together. By using both embodiments together, the compression rate of the foreground image data and the background image data is controlled according to the transparency of the image data that becomes the foreground when the images are superimposed, and the image deterioration is suppressed and the image Data compression rate can be increased.

  In addition, the said embodiment and modification can be suitably changed within the scope of the present invention.

  For example, in the above-described embodiment and modification, the predetermined quantization coefficient set in advance in the initial coefficient setting unit 25 is quantized by calculation using the correction values set in the pixel information determination unit 22 and the α value determination unit 23. The quantization coefficient is calculated, but a large number of quantization coefficient tables corresponding to one or both of the state of the matrix pattern of the pixel information determination unit 22 and the transparency of the α value determination unit 23 are stored in advance, and the state of the matrix pattern It is also possible to configure such that an appropriate table is selected and used according to the transparency.

  In the embodiment and the modification, the example in which the code generation unit 13 includes the variable length encoding unit 17 has been described. However, the code generation unit 13 is not particularly limited, and the variable length encoding unit 17 may not be included. In this case, the quantized data generated by the quantizing unit 16 can be used as encoded data.

  It is also possible to provide a means for compressing and encoding image data by other methods before and after the quantization unit 16. For example, a plurality of types of compression methods can be provided in series and sequentially compressed, and finally compressed data can be used as quantized data.

  Furthermore, in the above-described embodiment and modification, DCT (discrete cosine transform), which is one of transform coding methods, has been described as an example of “predetermined coding method” of image data. In addition to using the wavelet transform and Hadamard transform which are the above-described methods, it is also possible to use predictive coding using DPCM (Differential Pulse Code Modulation). Further, although the α encoder uses the same encoding algorithm as that of the Y, U, and V encoders, the α data may be encoded by an algorithm different from that of the Y, U, and V data. Furthermore, the present invention may be applied to any image data encoding method other than the above-described image data encoding method and any α encoder using a method other than the above-described α data encoding method.

  In the above embodiment, the transparency information is set for the whole of one image. However, the present invention can be applied to an image in which the transparency information is set for a part of the region. Is possible.

  In the above embodiment, an example of a moving image in which one image and the other image are superimposed and synthesized has been described. However, the number of images to be superimposed is not limited, and three or more layers of images are superimposed. The present invention may be applied to a moving image that configures each frame by combining them. Further, the present invention can be applied not only to encoding moving image data as in the above embodiment, but also to encoding still image data using alpha blending.

  The above embodiment is an exemplification of the present invention, and it is needless to say that the present invention is not limited to the above embodiment.

10: Encoding device (encoding device for image data)
11, 11y, 11u, 11v, 11α ... encoder 12 ... DCT unit 13 ... code generation unit 14 ... coefficient setting unit 15 ... prediction processing unit 16 ... quantization unit 17 ... variable length coding unit 21 ... memory unit 22 ... pixel information Determination unit 23 ... α value determination unit 24 ... coefficient control unit 25 ... initial coefficient setting unit 28 ... inverse quantization unit 29 ... inverse DCT unit 31 ... frame memory unit 32 ... prediction process execution unit

Claims (2)

An image data encoding device that compresses and encodes color information data in which transparency information is set for each pixel and transparency data as the transparency information data ,
A color information encoder that encodes color information data; and a transparency data encoder that encodes transparency data.
The color information encoder uses a predetermined quantization coefficient from an initial coefficient setting unit to determine a transparency reference correction number according to the transparency of the corresponding pixel and a pattern reference correction number according to the state of the pattern of the matrix including the pixel. A coefficient control unit that calculates a quantization coefficient by performing a predetermined operation on the signal and determines a divisor to be given to the quantization unit;
The coefficient control unit
The divisor is set to double when the opacity of the pixel set based on the transparency information is halved, and
An image data encoding apparatus configured to perform setting so that the transparency reference correction number is constant when the transparency of the pixel is equal to or greater than a predetermined value . The numerical value range of the transparency reference correction number used for the predetermined calculation in the coefficient control unit is set wider than the numerical value range of the design reference correction number, The size of the correction range of the predetermined quantization coefficient by the transparency reference correction number is set larger than the size of the correction range of the default quantization coefficient by the design reference correction number. The image data encoding device according to 1.

Images