JP2008236090A - Color space conversion apparatus and method, information processor, computer program, and information recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a color space conversion apparatus for applying a function for realizing reversible color space conversion, which is capable of generating a plurality of different color vectors from one color vector by changing the parameter of the function, a color space conversion method, an information processor, a computer program and an information recording medium. <P>SOLUTION: The color space conversion apparatus for performing reversible conversion between an RGB color space and a YUV color space to the color vector indicating a color includes a color space conversion means for generating Y, U and V which are the components of the color vector after the conversion with R, G and B as the components of the color vector before the conversion for the conversion from the RGB color space to the YUV color space. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a color space conversion device, a color space conversion method, an information processing device, a computer program, and an information recording medium.

  Conventionally, there is a color space conversion technique for converting a color space expressing a color. Such a technique is used, for example, when converting an RGB color space that is generally used in a computer or the like into a YUV color space suitable for compressing image data.

  By the way, since color images used for professional use and medical use require high quality, such images require high resolution, high image quality, etc., and the image data is compressed. In this case, it is desirable to apply a reversible compression method. Therefore, a reversible conversion method is also required for the color space conversion performed in the compression process. The reversible color space conversion is not limited to compression processing, and is a useful technique when a high-quality image is required.

Therefore, for example, in the invention such as the data compression device disclosed in Japanese Patent No. 3753782 (Patent Document 1), when performing color conversion, the forward conversion unit or the inverse conversion unit performs the conversion for each component of a plurality of color data vectors. It is disclosed that reversible color conversion is performed by using a predetermined mathematical expression. Note that “color conversion” in Patent Document 1 corresponds to “color space conversion” of the present application.
Japanese Patent No. 3753782

  However, the above Patent Document 1 only discloses several sets of mathematical expressions that realize reversible color space conversion. In each expression, one color vector is generated from one color vector by color space conversion. Because they are limited, these equations are just singularities on the function. Since the use of color images is diverse and tends to increase further, a reversible color space conversion technique suitable for each application is required. However, the reversible color space conversion corresponding to each application cannot be realized only by the equation disclosed in Patent Document 1.

  The present invention has been invented in order to solve these problems in view of the above points, and is a function that realizes reversible color space conversion. Since it was discovered that a plurality of different color vectors can be generated from the color vector of the color space, a color space conversion device to which the color vector is applied, a color space conversion method, an information processing device using the color space conversion method, a computer program, And it aims at providing an information recording medium.

  In order to achieve the above object, the color space conversion apparatus of the present invention employs the following configuration.

  A color space conversion apparatus according to the present invention is a color space conversion apparatus that performs reversible conversion between a RGB color space and a YUV color space for a color vector representing a color, and converts the RGB color space to a YUV color space. Is a color space conversion means for generating Y, U, and V, which are color vector components after conversion, based on the following equation (1), where R, G, and B are color vector components before conversion: It can be.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  Thus, there is provided a color space conversion device that realizes reversible color space conversion, and that applies a function that generates a plurality of different color vectors from one color vector by changing parameters of the function. be able to.

  In order to achieve the above object, a color space conversion device of the present invention is a color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, Conversion from the RGB color space to the YUV color space is based on the following equation (2), with the color vector components before conversion as R, G, and B, and the color vector components after conversion as Y, U, and V: The color space conversion means can be configured to be generated.

但し、α、β、γは、整数であって、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  Thus, there is provided a color space conversion device that realizes reversible color space conversion, and that applies a function that generates a plurality of different color vectors from one color vector by changing parameters of the function. be able to.

  In order to achieve the above object, a color space conversion device of the present invention is a color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, Conversion from the RGB color space to the YUV color space is based on the following equation (3), where R, G, and B are the color vector components before conversion, and Y, U, and V are the color vector components after conversion. The color space conversion means can be configured to be generated.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋γ＝β、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  Thus, there is provided a color space conversion device that realizes reversible color space conversion, and that applies a function that generates a plurality of different color vectors from one color vector by changing parameters of the function. be able to.

  In order to achieve the above object, a color space conversion device of the present invention is a color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, The conversion from the YUV color space to the RGB color space is based on the following equation (4), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. The color space conversion means can be configured to be generated.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  Thus, there is provided a color space conversion device that realizes reversible color space conversion, and that applies a function that generates a plurality of different color vectors from one color vector by changing parameters of the function. be able to.

  In order to achieve the above object, a color space conversion device of the present invention is a color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, The conversion from the YUV color space to the RGB color space is based on the following equation (5), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. The color space conversion means can be configured to be generated.

但し、α、β、γは、整数であって、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  Thus, there is provided a color space conversion device that realizes reversible color space conversion, and that applies a function that generates a plurality of different color vectors from one color vector by changing parameters of the function. be able to.

  In order to achieve the above object, a color space conversion device of the present invention is a color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, The conversion from the YUV color space to the RGB color space is based on the following equation (6), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. The color space conversion means can be configured to be generated.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋γ＝β、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  Thus, there is provided a color space conversion device that realizes reversible color space conversion, and that applies a function that generates a plurality of different color vectors from one color vector by changing parameters of the function. be able to.

  In order to achieve the above object, the color space conversion device of the present invention can be further configured such that (α + β + γ) is a power of 2.

  Thus, the configuration can be simplified by executing the division by a shift operation.

  In order to achieve the above object, the color space conversion method of the present invention is a color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, Conversion from the RGB color space to the YUV color space is based on the following equation (1), where R, G, B are the components of the color vector before conversion, and Y, U, V are the components of the color vector after conversion. The color space conversion step to be generated can be configured.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  In order to achieve the above object, the color space conversion method of the present invention is a color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, Conversion from the RGB color space to the YUV color space is based on the following equation (2), with the color vector components before conversion as R, G, and B, and the color vector components after conversion as Y, U, and V: The color space conversion step to be generated can be configured.

但し、α、β、γは、整数であって、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  In order to achieve the above object, the color space conversion method of the present invention is a color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, Conversion from the RGB color space to the YUV color space is based on the following equation (3), where R, G, and B are the color vector components before conversion, and Y, U, and V are the color vector components after conversion. The color space conversion step to be generated can be configured.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋γ＝β、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  In order to achieve the above object, the color space conversion method of the present invention is a color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, The conversion from the YUV color space to the RGB color space is based on the following equation (4), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. The color space conversion step to be generated can be configured.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  In order to achieve the above object, the color space conversion method of the present invention is a color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, The conversion from the YUV color space to the RGB color space is based on the following equation (5), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. The color space conversion step to be generated can be configured.

但し、α、β、γは、整数であって、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  In order to achieve the above object, the color space conversion method of the present invention is a color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color, The conversion from the YUV color space to the RGB color space is based on the following equation (6), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. The color space conversion step to be generated can be configured.

但し、α、β、γは、整数であって、α≠１、β≠２、γ≠１、α＋γ＝β、α＋β＋γ≠０、
Ｉ（ｘ）は、ｘの小数点以下を正の無限大方向に切り上げる又は負の無限大方向に切り捨てることにより整数化する関数、
である。

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.

  In order to achieve the above object, the color space conversion method of the present invention can be further configured such that (α + β + γ) is a power of 2.

  In order to achieve the above object, the information processing apparatus according to the present invention performs color space conversion for performing color space conversion for each pixel of image data using the color space conversion method according to claim 8. And compression means for compressing the image data that has been color space converted by the color space conversion means.

  In order to achieve the above object, an information processing apparatus according to the present invention includes an expansion unit that expands image data compressed by the information processing apparatus according to claim 15 and a pixel of the image data expanded by the expansion unit. A color space conversion unit that performs color space conversion by the color space conversion method according to any one of claims 8 to 14 can be provided.

  In order to achieve the above object, the present invention may further be a computer program for causing a computer to execute the color space conversion method, or a computer-readable information recording medium on which the computer program is recorded.

  According to the space conversion device, the color space conversion method, the information processing apparatus using the color space conversion method, the computer program, and the information recording medium storing the computer program according to the present invention, a function for realizing reversible color space conversion A color space conversion device, a color space conversion method, an information processing device, and a computer program that apply a function capable of generating a plurality of different color vectors from one color vector by changing the parameters of the function It is possible to provide an information recording medium.

(Description of general processing of color space conversion)
In order to facilitate understanding of the embodiment of the present invention, prior to the description of the embodiment of the present invention, a general color space conversion process and the color space conversion process are performed, and further an image encoding process is performed. Alternatively, an information processing apparatus that performs image decoding processing will be described.

  Color space conversion can be represented by a matrix. Therefore, for example, a conversion formula from the RGB color space to the YUV color space is represented by the following formula (7).

可逆な色空間変換では、式（７）の変換行列Ｋが、さらに、次の２つの条件を満たす必要がある。

In the reversible color space conversion, the conversion matrix K in Expression (7) needs to satisfy the following two conditions.

Condition 1: k ₃ + k ₄ + k ₅ = k ₆ + k ₇ + k ₈ = 0
Condition 2: k ₀ + k ₁ + k ₂ = 1
Condition 1 is a condition for setting the values of U and V, which are color differences, to 0 when R = G = B, that is, when the color is “gray” represented only by luminance. This is a condition for maintaining the effective number of bits of luminance.

  Note that “reversible color space conversion” refers to returning to the original color vector when color space conversion is performed on the color vector and the obtained color vector is subjected to reverse color space conversion. Refers to color space conversion.

(Example of information processing apparatus that performs conventional image encoding processing and image decoding processing)
FIG. 1A is a diagram illustrating an example of a functional configuration of an information processing apparatus that performs a conventional encoding process on an image. The image encoding process is a process for generating code data by compressing image data, and the image decoding process is for decompressing the code data generated by the compression to restore the image. It is processing.

  The information processing apparatus 1a in FIG. 1A includes preprocessing means 101, orthogonal transform means 102, quantization means 103, variable length encoding means 104, and code data generation means 105.

  The preprocessing unit 101 converts input image data into data suitable for image encoding. For example, the preprocessing unit 101 performs dynamic range change or level shift for each component of a color vector representing a color corresponding to each pixel of image data. Level shift is to subtract or add a predetermined value to each component of the color vector.

  The preprocessing unit 101 may also perform color space conversion on the color vector. The color space conversion is, for example, conversion from an RGB color space to a color space represented by luminance and color difference. By converting to a color space represented by luminance and color difference, a color vector expressed by a color space suitable for encoding rather than the RGB color space can be obtained. The color space represented by the luminance and the color difference is, for example, a YUV color space, a YCbCr color space, or a YCgCo color space.

  In the following embodiments, the YUV color space is treated as a general term for color spaces represented by luminance and color difference. That is, in the following embodiment, U and V may be read as Cb, Cr, Cg, Co, or the like.

  The orthogonal transform means 102 performs processing of dividing the image data into a plurality of stages from a visually important low frequency component to a visually less important high frequency component by decomposing the image data into frequency components. The frequency component obtained by this processing is called “coefficient data”.

  The quantizing unit 103 performs processing for changing the dynamic range of the coefficient data by dividing the coefficient data obtained by the orthogonal transform unit 102 by a predetermined quantization step width. This process is referred to as “quantization”, and the fractional value that has been subjected to processing such as truncation by division is not reconstructed during the decoding process, resulting in degradation of subjective image quality with respect to the decoded image. Therefore, in order to reduce the deterioration of subjective image quality as much as possible, low-frequency components that have a large effect on subjective image quality are not quantized as much as possible in the coefficient data, and high-frequency components that do not have a significant effect on subjective image quality are Divide by a value greater than the low frequency component. As a result, it is possible to reduce the amount of code data generated while preventing deterioration in subjective image quality.

  The variable length coding unit 104 performs variable length coding on the coefficient data quantized by the quantization unit 103 to reduce the amount of data. The variable length encoding unit 104 generates a variable length code by, for example, a Huffman encoding process or an arithmetic encoding process.

  The code data generation unit 105 generates code data including a packet or the like that is a group of codes obtained by adding predetermined header information or the like to the variable length code generated by the variable length encoding unit 104 or the like. The code data is called, for example, “code stream”, “bit stream”, or “byte stream”.

  FIG. 1B is a diagram illustrating an example of a functional configuration of an information processing apparatus that performs a conventional image decoding process. The information processing apparatus 1b in FIG. 1B includes code data analysis means 106, variable length decoding means 107, inverse quantization means 108, inverse orthogonal transform means 109, and post-processing means 110.

  For example, the code data analysis unit 106 analyzes the structure of a packet or the like included in the code data based on a marker or header information included in the code data, and outputs a variable-length code. The variable length decoding unit 107 decodes the variable length code output by the code data analysis unit 106. The variable length decoding means 107 decodes the variable length code produced | generated by the algorithm of the Huffman code or the arithmetic code, for example according to the algorithm.

  The inverse quantization unit 108 performs an inverse quantization process on the quantized coefficient data among the data decoded by the variable length decoding unit 107. Thereby, the dynamic range of coefficient data becomes the same as the dynamic range before quantization.

  The inverse orthogonal transform unit 109 performs an inverse orthogonal transform process on the coefficient data output from the inverse quantization unit 108. By this processing, image data corresponding to the image data before orthogonal transformation is output. The post-processing unit 110 performs reverse level shift processing or reverse color space conversion processing corresponding to the processing performed by the pre-processing unit 101.

  The information processing device 1a and the information processing device 1b may be configured as one device. In that case, for example, the orthogonal transform unit 102 and the inverse orthogonal transform unit 109 include the same processing, and thus are configured as one unit, thereby simplifying the configuration of the information processing apparatus. . Similarly, the quantization means 103 and the inverse quantization means 108, the variable length encoding means 104 and the variable length decoding means 107, and the preprocessing means 101 and the postprocessing means 110 are configured as one means respectively. The configuration of the information processing apparatus can be further simplified.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Note that the Y, U, and V components in Equations (1) to (6) may be read as Y, Cb, and Cr components or Y, Co, and Cg components, respectively.
[First embodiment]
The color space conversion device according to the first embodiment of the present invention has color space conversion means for converting color data from an RGB color space to a YUV color space, for example, according to the equation (1). It can be easily understood that the expression (1) satisfies the conditions 1 and 2 by expressing the conversion matrix K for realizing the color space conversion of the expression (1) by the following expression (8).

式（１）において、α、β、及び、γと、Ｒ、Ｇ、及び、Ｂとは、整数であるため、（α＋β＋γ）と（αＲ＋βＧ＋γＢ）とは整数となる。そこで、整数を整数で割った値である（αＲ＋βＧ＋γＢ）／（α＋β＋γ）＝εとすると、εは実数となる。

In the formula (1), α, β, and γ and R, G, and B are integers, so (α + β + γ) and (αR + βG + γB) are integers. Therefore, if (αR + βG + γB) / (α + β + γ) = ε, which is a value obtained by dividing the integer by the integer, ε is a real number.

  By the way, in the calculation means using a computer or the like, an error occurs in the value of the part after the decimal point of the real number depending on the number of significant digits. However, no error occurs for the integer part. Therefore, the floor function value that is the maximum integer value not exceeding ε is uniquely determined. Similarly, the sealing function value which is the smallest integer value not less than ε is also uniquely determined.

  The floor function is a function that rounds down the decimal point of the real number in the negative infinity direction, and the ceiling function is a function that rounds up the decimal point of the real number in the positive infinity direction. That is, the floor function can be realized by a right shift operation, and the ceiling function can be realized by a right shift operation on a value obtained by adding 1 to a real number to which the ceiling function is applied.

  That is, in equation (1), when any one of the floor function and the ceiling function is used, no error occurs in the Y value regardless of the number of significant digits of the calculation means. Further, U and V, which are differences between integers, do not cause an error.

  Next, with respect to reversibility when the color data in the YUV color space obtained by the color space conversion according to the equation (1) is converted into the color data of the RGB color space by the color space conversion according to the equation (4), the following equation (9 ) To Equation (11).

  The following equation (9) can be derived from the equation of Y in equation (1).

さらに、式（９）より次式（１０）を導くことができる。

Further, the following equation (10) can be derived from the equation (9).

式（１０）において、Ｙ、Ｕ、Ｖは式（１）によって得られた誤差のない整数であり、Ｉ｛（αＵ＋γＶ）／（α＋β＋γ）｝も、Ｉ（ｘ）が、フロア関数又はシーリング関数であるために誤差のない整数である。したがって、誤差のない整数の和であるＧも誤差のない整数となる。さらに、式（１）より、次式（１１）が得られる。

In the equation (10), Y, U, and V are integers without error obtained by the equation (1), and I {(αU + γV) / (α + β + γ)} is equal to I (x) is a floor function or a ceiling function. Therefore, it is an integer without error. Therefore, G, which is the sum of integers without error, is also an integer without error. Further, the following equation (11) is obtained from the equation (1).

式（１１）において、Ｕ、Ｖ、Ｇが誤差のない整数であるため、Ｒ及びＢも誤差のない整数となる。

In Equation (11), U, V, and G are integers without errors, so R and B are also integers without errors.

  From Equation (10) and Equation (11), it can be seen that Equation (4) provides R, G, and B values without error.

  Therefore, the color space conversion based on Expression (1) is reversible, and the inverse conversion is represented by Expression (4). Moreover, it can be easily understood that the color space conversion based on the equation (4) is reversible and the inverse conversion is expressed by the equation (1), as described above.

In addition, when color space conversion is performed using I (x) as a floor function in Expression (1), the floor function is also used in Expression (4) which is the inverse conversion. Similarly, when color space conversion is performed using I (x) as a sealing function in Expression (1), the ceiling function is also used in Expression (4) which is the inverse conversion. The same applies to the case where the expression (4) is a positive conversion.
[Second Embodiment]
The color space conversion apparatus according to the second embodiment of the present invention has color space conversion means for converting color data from the RGB color space to the YUV color space, for example, according to the equation (2). The fact that Expression (2) satisfies Condition 1 and Condition 2 can be easily understood by expressing the color space conversion of Expression (2) by a conversion matrix, and thus description thereof is omitted here.

  Furthermore, in the formula (2), α, β, and γ, and R, G, and B are integers. Therefore, as in the formula (1), either the floor function or the ceiling function is selected. When the function is used, no error occurs in the Y value regardless of the number of significant digits of the calculation means. Further, U and V, which are differences between integers, do not cause an error.

  Next, the reversibility when the color data of the YUV color space obtained by the color space conversion according to the equation (2) is converted into the color data of the RGB color space by the color space conversion according to the equation (5) is expressed by the equation (12). And (13).

  Y in Formula (2) is the same as Y in Formula (1), and Formula (9) is obtained from Formula (2). Furthermore, the following equation (12) can be derived from the equation (9).

式（１２）における（αＵ＋（α＋γ）Ｖ）／（α＋β＋γ）は、整数同士の演算であり、その値は実数になるため、第一の実施の形態での説明と同様に、式（１２）におけるフロア関数値又はシーリング関数値は、誤差が生じない。したがって、誤差のない整数Ｙから誤差のない整数値を減じて得られるＧにも、誤差が生じない。

In equation (12), (αU + (α + γ) V) / (α + β + γ) is an arithmetic operation between integers, and the value is a real number. Therefore, as in the description in the first embodiment, equation (12) No error occurs in the floor function value or the ceiling function value at. Accordingly, no error occurs even in G obtained by subtracting an integer value having no error from the integer Y having no error.

  In the following equation (13) obtained from equation (2), V and G are integers without error, so B is also an integer without error. Further, since U and B are integers without error, R is also an integer without error.

式（１２）及び式（１３）とから、式（５）によって、誤差のないＲ、Ｇ、Ｂの値が得られることが判る。

From Equation (12) and Equation (13), it can be seen that Equation (5) provides R, G, and B values without error.

  Therefore, the color space conversion based on Expression (2) is reversible, and the inverse conversion is represented by Expression (5). Moreover, it can be easily understood that the color space conversion based on the equation (5) is reversible and the inverse conversion is expressed by the equation (2) as described above.

In addition, when color space conversion is performed using I (x) as a floor function in Expression (2), the floor function is also used in Expression (5) which is the inverse conversion. Similarly, when color space conversion is performed using I (x) as a sealing function in Expression (2), the ceiling function is also used in Expression (5) which is the inverse conversion. The same applies to the case where the expression (5) is a positive conversion.
[Third embodiment]
The color space conversion apparatus according to the third embodiment of the present invention has color space conversion means for converting color data from the RGB color space to the YUV color space, for example, according to the equation (3). The fact that the expression (3) satisfies the condition 1 can be easily understood by expressing the color space conversion of the expression (3) by a conversion matrix, and thus the description thereof is omitted here. Further, since the expression (3) satisfies the condition 2, the relationship of the following expression (14) is established.

さらに、式（３）において、α、β、及び、γと、Ｒ、Ｇ、及び、Ｂとは、整数であるため、式（３）と同様に、フロア関数又はシーリング関数の何れか一の関数を用いる場合に、Ｙ及びＶの値には計算手段の有効桁数によらず誤差が生じない。さらに、整数同士の差であるＵも誤差は生じない。

Furthermore, in Expression (3), α, β, and γ, and R, G, and B are integers. Therefore, as in Expression (3), either the floor function or the ceiling function is used. When the function is used, no error occurs in the values of Y and V regardless of the number of significant digits of the calculation means. Furthermore, U, which is a difference between integers, does not cause an error.

  Next, the reversibility in the case where the color data of the YUV color space obtained by the color space conversion according to the equation (3) is converted into the color data of the RGB color space by the color space conversion according to the equation (6) is expressed by the following equation (15). ) Etc.

  By substituting equation (14) for Y in equation (3), equation (15) can be derived. Further, the definition of Expression (16) is applied to Expression (15) to obtain Expression (17).

同様に、式（３）におけるＶに対し、式（１４）を代入し、さらに式（１６）の定義を適用して式（１８）及び（１９）を得る。

Similarly, Expression (14) is substituted for V in Expression (3), and Expressions (18) and (19) are obtained by applying the definition of Expression (16).

ところで、フロア関数又はシーリング関数には、式（２０）の関係が成り立つ。

By the way, the relationship of Expression (20) is established in the floor function or the sealing function.

そこで式（２０）を、式（１７）に代入することにより、式（２１）が導かれ、さらに、式（２２）が得られる。

Therefore, by substituting equation (20) into equation (17), equation (21) is derived, and further equation (22) is obtained.

さらに、式（１８）に、式（３）のＵ＝Ｒ−Ｂから導かれる式（２３）を代入することにより、式（２４）が導かれ、さらに式（２５）が得られる。

Further, by substituting equation (23) derived from U = RB in equation (3) into equation (18), equation (24) is derived, and further equation (25) is obtained.

式（２２）、式（２３）及び式（２５）から、誤差のないＹ、Ｕ、Ｖと、整数α、β、γとから、誤差のないＲ、Ｇ、Ｂの値が得られることが判る。

From equations (22), (23), and (25), there can be obtained error-free values of R, G, and B from Y, U, and V with no error and integers α, β, and γ. I understand.

  Therefore, the color space conversion based on Expression (3) is reversible, and the inverse conversion is represented by Expression (6). Further, it can be easily understood that the color space conversion based on the equation (6) is reversible and the inverse conversion is expressed by the equation (3) as described above.

In addition, when color space conversion is performed using I (x) as a floor function in Expression (3), the floor function is also used in Expression (6) which is the inverse conversion. Similarly, when color space conversion is performed using I (x) as a sealing function in Expression (3), the ceiling function is also used in Expression (6) which is the inverse conversion. The same applies to the case where the expression (6) is a positive conversion.
[Fourth embodiment]
FIG. 2 is a diagram for explaining an example of an image encoding process and an image decoding process in the information processing apparatus according to the embodiment of the present invention. The information processing apparatus 2a in FIG. 2A compresses input image data to generate code data, and the information processing apparatus 2b in FIG. 2B decodes the compressed code data to generate an image. To get. The process performed by the information processing apparatus 2a is a reversible encoding process, and the process performed by the information processing apparatus 2b is a process of decoding code data generated by the encoding process. The information processing device 2a and the information processing device 2b may be configured as one device.

  Note that “reversible encoding processing” means that when code data generated by encoding is decoded and an image is formed again, the image (hereinafter referred to as “decoded image”) is encoded. This is a process of generating code data that is the same as the previous image (hereinafter referred to as “original image”).

  The information processing apparatus 2a in FIG. 2 includes a DC level shift and color space conversion processing unit 201, a wavelet conversion processing unit 202, a bit plane encoding unit 204, a packet generation unit 205, and a code data generation unit 206.

  Since the information processing apparatus 2a is not provided with the quantization unit 203, there is no loss of image information due to quantization. The DC level shift process, wavelet transform process, bit plane encoding process, packet generation process, and code data generation process are all reversible processes, and the number of effective digits of the calculation means included in the information processing apparatus 2a is sufficient. If so, there is no loss of image information. Therefore, the information processing apparatus 2a can perform reversible encoding processing by changing the color space conversion processing to reversible color space conversion processing based on equations (1) to (3).

  Processing of wavelet transform processing unit 202, processing of bit plane encoding unit 204, processing of packet generation unit 205, processing of code data generation unit 206, and processing of DC level shift in DC level shift and color space conversion processing unit 201 Performs, for example, processing according to the JPEG2000 (ISO / IEC 15444-1) standard. As a result, the other processes except the color space conversion process can be made the same as the reversible encoding process in the JPEG2000 standard, and an information processing apparatus that performs the encoding process of the JPEG2000 standard and means for performing these processes are provided. Can be shared.

(Explanation of DC level shift and color space conversion)
The DC level shift and color space conversion processing unit 201 adds or subtracts the same value for all pixel values to the input image data, thereby maintaining the pixel value while maintaining the dynamic range. To change. Thereby, the compression process performed in the latter stage can be made more efficient.

  The conversion of the DC level shift is performed based on, for example, Expression (26).

P (x, y) = I (x, y) -2 ^ Ssiz (i) (26)
Where P (x, y) is the pixel value at the coordinate (x, y) after DC level shift conversion,
I (x, y) is a pixel value at coordinates (x, y) before DC level shift conversion,
i is a counter representing a component in the image data before conversion. For example, R is i = 0, G is i = 1, B is i = 2,
Ssiz (i) is the number obtained by subtracting 1 from the bit depth of each component,
2 ^ Ssiz (i) is 2 to the power of Ssiz (i),
It is.

  That is, a process of subtracting half the value of the dynamic range for each component from the value of each pixel is performed according to Expression (26). Note that the DC level shift is not applied to a component having a sign. The component having a sign is, for example, U and V in the YUV color space.

  The DC level shift and color space conversion processing unit 201 further converts the color vector representing the color into a color vector represented in the YUV color space when the input image data is image data represented in the RGB color space. Performs color space conversion processing. The color space conversion process performed here is based on, for example, any one of the expressions (1) to (3).

(Description of processing to set color space conversion parameters)
FIG. 3 is a diagram illustrating an example of a detailed functional configuration of the DC level shift and color space conversion processing unit 201. The DC level shift and color space conversion processing unit 201 in FIG. 3 includes a DC level shift processing unit 21, a color space conversion unit 22, and a parameter setting unit 23, for example. The DC level shift processing unit 21 changes the value of the pixel by adding or subtracting a predetermined value to the input image data based on, for example, Expression (26).

  The color space conversion unit 22 performs reversible color space conversion. For example, the color space conversion unit 22 performs color space conversion based on any one of the expressions (1) to (3). The color space conversion unit 22 may also perform color space conversion based on, for example, any one of Expressions (4) to (5). Thereby, reversible color space conversion can be realized.

  The parameter setting unit 23 is included in the equations (1) to (3) when the color space conversion unit 22 performs reversible color space conversion based on any one of the equations (1) to (3). Set the values of parameters α, β, and γ. The parameter setting unit 23 may set each value by a predetermined combination of α, β, and γ values. The parameter setting unit 23 may have a configuration that holds a plurality of combinations thereof.

  The parameter setting unit 23 may set the parameter value in accordance with, for example, the property of the image corresponding to the input image data. For example, the speed and accuracy of processing required for the color space conversion processing The parameter value may be set corresponding to the above.

  For example, the parameter setting unit 23 may be configured to set the values of α, β, and γ in a combination in which (α + β + γ) is a power of 2. As a result, the division in the color space conversion processing of the equations (1) to (6) can be realized by the shift operation, the processing can be performed at high speed, and the color space conversion can be performed by hardware such as a circuit. When configured with hardware, the scale of the circuit and the like can be reduced.

  Furthermore, in addition to (α + β + γ), by combining α, β, and γ with powers of 2, the processing can be performed at higher speed, and the scale of a hardware circuit or the like that realizes the processing Can be reduced.

  If the reversible color space conversion performed by the color space conversion unit 22 is based on a set of parameters, the parameter setting unit 23 may be omitted, and the color space conversion unit 22 is determined in advance. Color space conversion may be performed based on any one of the expressions (1) to (6) based on a set of parameters. Thereby, when the property of the image processed by the information processing apparatus 2a, the quality required for the code data, the required processing speed, and the like are limited, the configuration of the apparatus can be simplified.

(Explanation of wavelet transform formula)
The wavelet transform processing unit 202 performs wavelet transform on the image data processed by the DC level shift and color space conversion processing unit 201 and converts the image data into coefficient data for each frequency component. Examples of conversion equations for wavelet transformation performed by the wavelet transformation processing unit 202 are shown in Equation (27) and Equation (28).

[Step 1] C (2i + 1) = P (2i + 1) −floor ((P (2i) + P (2i + 2)) / 2) (27)
[Step 2] C (2i) = P (2i + 1) + floor ((C (2i-1) + C (2i + 1) +2) / 4) (28)
Where P (n) is the input value at position n before wavelet transformation,
floor (x) is a function that rounds off the fractional part of the value x in the direction of negative infinity,
It is.

  Expressions (27) and (28) are expressions of 5 × 3 wavelet transform, and an output of one low-pass filter is obtained using five adjacent pixels, and further, one expression is obtained using three adjacent pixels. Get the output of the high-pass filter. That is, C (2i) obtained by Expression (28) is the output of the low-pass filter, and C (2i + 1) obtained by Expression (27) is the output of the high-pass filter. The output of the low-pass filter is called “L coefficient”, and the output of the high-pass filter is called “H coefficient”.

(Description of wavelet transform process)
4 to 8 are diagrams for explaining the process of the two-dimensional wavelet transform in detail. FIG. 4 is an example of image data input to the wavelet transform processing unit 202. The image data of FIG. 4 corresponds to an image of 16 pixels in length and width, and is obtained by performing DC level shift on the pixel value.

  When two-dimensional wavelet transform is performed on the image data of FIG. 4, first, one-dimensional wavelet transform is performed on either the horizontal direction or the vertical direction, and then 1 is performed on the other. Perform dimension wavelet transform. When the JPEG2000 standard is followed, vertical wavelet transformation is performed first, and horizontal wavelet transformation is performed on the output.

  Therefore, in FIG. 4, the horizontal direction is the x direction, the vertical direction is the y direction, and the pixel value of the pixel represented by the coordinate y is P (y) at a certain coordinate x in the horizontal direction. However, 0 ≦ x ≦ 15 and 0 ≦ y ≦ 15.

  In order to perform the wavelet transform process in the vertical direction, first, high-pass filter processing according to Expression (28) is performed around the pixel P (y) whose y coordinate is an odd number (y = 2i + 1) at a certain coordinate x in the horizontal direction. . In FIG. 4, “odd” is written at the left end of the row having an odd y coordinate.

  Next, at the same coordinate x, the low-pass filter process according to Expression (27) is performed using the coefficient obtained by the high-pass filter process and P (y) around the pixel P (y) whose y-coordinate is an even number. . In FIG. 4, “even” is displayed at the left end of the row having an even y coordinate.

  With respect to all coordinates x (where 0 ≦ x ≦ 15), the above-described processing is performed for each of pixels having an odd y coordinate and pixels having an even y coordinate.

  Note that when performing processing based on Expression (27) and Expression (28) for pixels located near the edge of the image, virtual pixels are arranged outside the edge of the image by a predetermined method. The predetermined method is, for example, a method in which pixels outside the image are arranged so that the pixels are targeted at the inside of the image and the outside of the image around the edge pixel or the edge of the image.

  FIG. 5 is an example of an array of data after performing the wavelet transform process in the vertical direction for all coordinates x. In FIG. 5, the L coefficient is represented as “L” and the H coefficient is represented as “H”. In FIG. 5, “H” is arranged in a row having an odd y coordinate, and “L” is arranged in a row having an even y coordinate.

  Next, horizontal wavelet transform processing is performed on the data array of FIG. In the horizontal wavelet transform process, first, in the data array of FIG. 5, at a certain coordinate y in the vertical direction, the data P (x) whose x coordinate is an odd number (x = 2i + 1) is centered. Next, at the same coordinate y, the low-pass filter process according to the equation (27) is performed around the data P (x) whose x coordinate is an even number.

  FIG. 6 is an example of the data arrangement after the horizontal wavelet transform process is performed on the data of FIG. In FIG. 6, the LL coefficient, which is coefficient data obtained by performing low-pass filter processing in each of the vertical direction and the horizontal direction, is “LL”, obtained by performing high-pass filter processing in the vertical direction and low-pass filter processing in the horizontal direction. The coefficient data obtained by performing the low-pass filter process in the vertical direction and the high-pass filter process in the horizontal direction as the coefficient data obtained by performing the low-pass filter process in the vertical direction and the high-pass filter process in the horizontal direction as “LH”. The HH coefficient, which is coefficient data obtained by performing the high-pass filter process, is denoted as “HH”.

  As shown in FIG. 6, performing wavelet transform once in each of the vertical direction and the horizontal direction is referred to as “one-time decomposition” or “decomposition level 1”.

  FIG. 7 shows the data array shown in FIG. 6 rearranged by collecting the subbands of the LL coefficient, the HL coefficient, the LH coefficient, and the HH coefficient. The classification for each sub-band is called “deinterleaving”, and conversely, returning from the arrangement of FIG. 7 to the arrangement of FIG. 6 is called “interleaving”. A subband made up of LL coefficients is an image having a resolution half that of the original image. In FIG. 7, the notation “1” such as “1LL” indicates that the coefficient data is a coefficient of decomposition level 1.

  FIG. 8 is a diagram illustrating an array of data obtained as a result of performing wavelet transform processing in the vertical direction and horizontal direction on the subbands including the LL coefficients in FIG. 7 and further performing deinterleaving. . In FIG. 8, the location where the LL coefficient of the decomposition level 1 in FIG. 7 is arranged becomes the LL coefficient, the HL coefficient, the LH coefficient, and the HH coefficient of the decomposition level 2.

  FIG. 9 is a diagram for explaining the relationship between the decomposition level and the resolution. In FIG. 9, resolution levels corresponding to coefficient data of decomposition levels 1 to 3 are shown. In FIG. 9, the LL coefficient at the decomposition level 3 is associated with the resolution level 0, and the HL coefficient, the LH coefficient, and the HH coefficient at the decomposition level 3 are associated with the resolution level 1. The coefficient at the decomposition level 2 is associated with the resolution level 2, and the coefficient at the decomposition level 1 is associated with the resolution level 3.

  As is clear from FIG. 9, the numbers representing the decomposition levels and the numbers representing the resolution levels are almost reversed in descending order or ascending order, and the resolution level is the decomposition level corresponding only to the LL coefficient. One less value.

(Encoding process for each bit plane)
FIG. 10 is a diagram illustrating processing performed by the bit plane encoding unit 204. FIG. 10A shows, for example, a block of coefficient data included in the region of the LL coefficient at the decomposition level 2 in FIG. 8, and is a rectangular unit called a code block. In the process of encoding for each bit plane, first, the region of the LL coefficient is divided into code blocks, and encoding for each bit plane is performed for each code block.

  FIG. 10B is a diagram illustrating a bit plane corresponding to the code block in FIG. The bit plane is a plane composed of values where the bit positions of the coefficient data included in the code block are the same. FIG. 10B shows four bit planes from the most significant bit plane on the MSB side to the least significant bit plane on the LSB side.

  FIG. 10C is a diagram illustrating a sub bit plane when entropy encoding is performed on a value included in the bit plane. The bit plane encoding unit 204 scans the values included in one bit plane with a maximum of three coding passes. A plane composed of values scanned by each coding pass is called a sub-bit plane. A coding pass is simply referred to as a “pass”.

  There are three paths: a signature propagation pass, a magnesium refinement pass, and a cleanup pass. Significance propagation path is a path that encodes non-significant coefficients with significant coefficients around them, and significant refinement pass is a path that encodes significant coefficients, and cleanup pass corresponds to the above two paths. This is a pass for encoding the remaining bits that are not.

  In FIG. 10C, a bit corresponding to the signature propagation pass is represented by 'p', a bit corresponding to the magnesium refinement pass is represented by 'r', and a bit corresponding to the cleanup pass is represented by 'c'.

  Note that each bit plane does not necessarily have all sub-bit planes corresponding to the above three paths, and some bit planes have only one or two sub-bit planes. Further, the most significant bit plane on the MSB side is always only the cleanup pass.

(Encoding process for each bit plane)
FIG. 11 is a table for explaining an example of the order of codes generated by the bit-plane encoding unit 204. The codes are arranged in order from the upper subband level, and each subband is subordinate to the upper subbitplane. The bit planes are arranged in order. In the bit plane encoding unit 204, for example, a code is generated by an algorithm of MQ encoding.

(Packet generation processing and code data generation processing)
The packet generation unit 205 sets the code generated by the bit plane encoding unit 204 to, for example, a packet for each sub bit plane shown in FIG. The code data generation unit 206 arranges the packets generated by the packet generation unit 205 in a predetermined order and further adds header information and the like to generate code data.

(Description of decryption process)
The code data generated by the code data generation unit 205 is decoded by the information processing apparatus 2b, and an image is output. 2B includes a packet acquisition unit 216, a packet analysis unit 215, a bit plane decoding unit 214, an inverse wavelet transform processing unit 212, and an inverse DC level shift and inverse color space conversion processing unit 211. Have.

  The packet acquisition unit 216 extracts a packet from the code data based on a marker or the like. The packet analysis unit 215 extracts a code from the packet. The bit plane decoding unit 214 decodes the extracted code and obtains coefficient data. The inverse wavelet transform processing unit 212 performs an inverse wavelet transform process on the coefficient data.

  The inverse wavelet transform process is executed based on the following equations (29) and (30).

[Step 1] P (2i) = C (2i) −floor (C (2i−1) + C (2i + 1) +2) / 4) (29)
[Step 2] P (2i + 1) = C (2i + 1) + floor ((P (2i) + P (2i + 2)) / 2) (30)
Where C (n) is the input value at position n before the inverse wavelet transform,
floor (x) is a function that rounds off the fractional part of the value x in the direction of negative infinity,
It is.

  The inverse DC level shift and inverse color space conversion processing unit 211 performs an inverse color space conversion process and an inverse DC level shift process to decode an image. In the inverse color space conversion process, for example, color space conversion is performed based on any one of Expressions (1) to (6) corresponding to the color space conversion at the time of encoding. . That is, when color space conversion is performed based on Expression (1) during encoding, inverse color space conversion based on Expression (4) is performed. Further, when color space conversion is performed based on Expression (2) at the time of encoding, inverse color space conversion based on Expression (5) is performed. In addition, when color space conversion is performed based on Expression (3) at the time of encoding, inverse color space conversion based on Expression (6) is performed.

  In addition, when color space conversion based on equations (4) to (6) is performed at the time of encoding, inverse color space conversion based on equations (1) to (3) is performed, respectively.

  The inverse color space conversion process is also performed with the same parameter value as that used in the color space conversion process, for example. That is, the values of α, β, and γ are set to be the same as those used in the color space conversion process. In order to realize this, the code data generation unit 206 may include a parameter value in the code data, for example. Further, instead of the parameter value, information representing the parameter value may be included.

  Thereby, it is possible to decode an image composed of pixels having the same color value as the image before encoding.

  The inverse DC level shift is executed based on the following equation (31), for example.

I (x, y) = P (x, y) -2 ^ Ssiz (i) (31)
However, I (x, y) is a pixel value at coordinates (x, y) after inverse DC level shift conversion,
P (x, y) is the value of the pixel at the coordinates (x, y) before the inverse DC level shift conversion,
i is a counter representing a component in the image data before conversion, for example, Y is i = 0, U is i = 1, V is i = 2,
Ssiz (i) is the number obtained by subtracting 1 from the bit depth of each component,
2 ^ Ssiz (i) is 2 to the power of Ssiz (i),
It is.

(Example of information processing apparatus based on MPEG-4 AVC / H.264 standard)
FIG. 12 is a diagram for explaining an example of image encoding processing in the information processing apparatus according to the embodiment of the present invention. For example, MPEG-4 AVC (ISO / IEC 14496-10) / ITU-TH . 2 is an example of an information processing apparatus that performs an image encoding process or an image decoding process based on the H.264 (hereinafter referred to as “AVC / H.264”) standard.

  The information processing apparatus 3a in FIG. 12A includes a color space conversion processing unit 301, a DCT / wavelet conversion processing unit 302, a CAVLC / CABAC processing unit 304, a NAL generation unit 305, and a byte stream format generation unit 306. .

  The color space conversion processing unit 301 performs reversible color space conversion processing based on, for example, Expressions (1) to (3) for input image data. The DCT / wavelet transform processing unit 302 performs orthogonal transform such as DCT or wavelet transform on the input data, and outputs coefficient data. The CAVLC / CABAC processing unit 304 performs variable length coding processing on the coefficient data. The variable length encoding process executed by the CAVLC / CABAC processing unit 304 is, for example, AVC / H. 264, a context adaptive variable length coding scheme (hereinafter referred to as “CAVLC”) or AVC / H.264 standard. This is a variable length coding process according to the context adaptive binary arithmetic coding system (hereinafter referred to as “CABAC”) according to the H.264 standard.

  The NAL generation unit 305 generates an AVC / H. A network abstraction layer (hereinafter referred to as “NAL”) according to the H.264 standard is generated. The byte stream format generation unit 306 adds a predetermined start code to the NAL generated by the NAL generation unit to generate code data in the byte stream format.

  Through the above processing, reversible image encoding processing can be realized by the information processing apparatus 3a.

  Also, the processing of the DCT / wavelet transform processing unit 302, the processing of the CAVLC / CABAC processing unit 304, the processing of the NAL generation unit 305, and the processing of the byte stream format generation unit 306 are performed in accordance with AVC / H. By performing processing according to the H.264 standard, other processing excluding color space conversion can be performed using AVC / H. H.264 standard, AVC / H. The information processing apparatus that generates code data according to the H.264 standard and the means for performing the image encoding process can be shared.

  The encoded data generated by the information processing device 3a is decoded by a procedure reverse to the generation to obtain a decoded image.

  FIG. 12B is a diagram illustrating the information processing apparatus 3a that performs a process of decoding the generated code data. The information processing apparatus 3 a includes a NAL acquisition unit 316, a NAL analysis unit 315, a CAVLD / CABAD processing unit 314, an inverse DCT / inverse wavelet transformation processing unit 312, and an inverse color space transformation processing unit 311.

  The NAL acquisition unit 316 acquires a NAL from the byte stream format based on a predetermined start code. The NAL analyzing unit 315 acquires a variable length code and the like from the NAL.

  The CAVLD / CABAC processing unit 314 performs variable length decoding processing on the variable length code to obtain coefficient data. The inverse DCT / inverse wavelet transform processing unit 312 performs inverse orthogonal transform on the coefficient data.

  The inverse color space conversion processing unit 311 obtains a decoded image by performing color space conversion on the pixel values obtained by performing the inverse orthogonal transform.

  Note that the inverse color space conversion process is performed based on, for example, any one of the equations (4) to (6). More specifically, when the color space conversion process is based on Expression (1), the inverse color space conversion based on Expression (4) is performed, and the color space conversion process is performed based on Expression (2). If the color space conversion process is based on Expression (3), the inverse color space conversion based on Expression (6) is performed. .

  With the above processing, the information decoding apparatus 3b can realize the image decoding processing for decoding the code data generated by the reversible image encoding processing and obtaining the same decoded image as the original image.

  In addition, the processing of the inverse DCT / inverse wavelet transform processing unit 312, the processing of the CAVLD / CABAD processing unit 314, the processing of the NAL analysis unit 315, and the processing of the NAL acquisition unit 316 are performed using the AVC / H. By performing processing according to the H.264 standard, other processing excluding color space conversion can be performed using AVC / H. H.264 standard, AVC / H. An information processing apparatus that decodes code data according to the H.264 standard and a means for performing image decoding processing can be shared.

(Example of image encoding process and image decoding process by information processing apparatus according to this embodiment)
FIGS. 13 to 18 are flowcharts of examples of image encoding processing and image decoding processing by the information processing apparatus according to the embodiment of the present invention. FIG. 13 is an example of an image encoding process including a color space conversion process based on Expression (1), and FIG. 14 is an example of an image decoding process including an inverse color space conversion process based on Expression (4). FIG. 15 is an example of an image encoding process including a color space conversion process based on Expression (2), and FIG. 16 is an example of an image decoding process including an inverse color space conversion process based on Expression (5). is there. FIG. 17 is an example of an image encoding process including a color space conversion process based on Expression (3), and FIG. 18 is an example of an image decoding process including an inverse color space conversion process based on Expression (6). is there.

  FIG. 13 is an example of processing in which code data a is generated from image data by the information processing apparatus 2a, for example. In step S101 of FIG. 13, image data represented in the RGB color space is input to the information processing apparatus 2a. Progressing to step S102 following step S101, the DC level shift and color space conversion processing unit 201 performs DC level shift based on Expression (26) on the input image data. Progressing to step S103 following step S102, the DC level shift and color space conversion processing unit 201 performs color space conversion on the image data on which the DC level shift has been performed, based on Expression (1). More specifically, for example, the color space conversion unit 22 performs color space conversion processing.

  Progressing to step S104 following step S103, the wavelet transform processing unit 202 performs 5 × 3 wavelet transform processing on the image data that has undergone color space transformation based on the equations (27) and (28). Output coefficient data. Progressing to step S105 following step S104, the bit plane encoding unit 204 performs MQ encoding for each bit plane on the coefficient data output in step S105.

  Progressing to step S106 following step S105, the packet generation unit 205 generates a packet from the MQ code or the like generated in step S105. Progressing to step S107 following step S106, the code data generation unit 206 generates code data a including the packet generated in step S106.

  FIG. 14 is a flowchart of an example of a process in which the information processing device 2b obtains a decoded image by decompressing the code data a generated by the process of FIG. In step S201 in FIG. 14, the code data a generated by the process in FIG. 13 is input to the information processing apparatus 2b. Progressing to step S202 following step S201, the packet acquisition unit 215 acquires a packet from the code data a. Progressing to step S203 following step S202, the packet analysis unit 215 acquires an MQ code from the packet acquired in step S202, and the bit plane decoding unit 214 further decodes the MQ code and outputs coefficient data. .

  Progressing to step S204 following step S203, the inverse wavelet transform processing unit 212 performs inverse wavelet transform based on the equations (29) and (30) on the coefficient data output in step S203. Progressing to step S205 following step S204, the inverse DC level shift and inverse color space conversion processing unit 211 performs an inverse color space conversion process based on Expression (4).

  Proceeding to step S206 following step S205, the inverse DC level shift and inverse color space conversion processing unit 211 performs an inverse DC level shift process on the image data on which the inverse color space conversion process has been performed, so that the RGB color space The decoded image represented by is acquired.

  FIG. 15 is an example of processing in which code data b is generated from image data by the information processing apparatus 2a, for example. The processing in step S301 and step S302 in FIG. 15 is the same as the processing in step S101 and step S102 in FIG. 13, and the processing from step S304 to step S307 in FIG. 15 is the same as the processing from step S104 to step S107 in FIG. Since it is the same, description is abbreviate | omitted here.

  In step S303 in FIG. 15, the DC level shift and color space conversion processing unit 201 performs color space conversion on the image data on which the DC level shift has been performed based on Expression (2). More specifically, for example, the color space conversion unit 22 performs color space conversion processing. Thereby, a reversible encoding process can be performed.

  FIG. 16 is a flowchart of an example of a process in which the information processing device 2b obtains a decoded image by decompressing the code data b generated by the process of FIG. The processing from step S401 to step S404 in FIG. 16 is the same as the processing from step S201 to step S204 in FIG. 14, and the processing in step S406 and step S407 in FIG. 16 is the same as the processing in step S206 and step S207 in FIG. Since it is the same, description is abbreviate | omitted here.

  In step S405, the inverse DC level shift and inverse color space conversion processing unit 211 performs an inverse color space conversion process based on Expression (5).

  Through the above processing, the encoded data generated by the reversible image encoding processing can be decoded, and a decoded image represented in the RGB color space can be acquired.

  FIG. 17 is an example of processing in which code data b is generated from image data by the information processing apparatus 2a, for example. The processing in step S501 and step S502 in FIG. 17 is the same as the processing in step S101 and step S102 in FIG. 13, and the processing in step S504 to step S507 in FIG. 17 is the same as the processing in step S104 to step S107 in FIG. Since it is the same, description is abbreviate | omitted here.

  In step S503 in FIG. 17, the DC level shift and color space conversion processing unit 201 performs color space conversion on the image data on which the DC level shift has been performed based on Expression (3). More specifically, for example, the color space conversion unit 22 performs color space conversion processing. Thereby, a reversible encoding process can be performed.

  FIG. 18 is a flowchart of an example of a process in which the information processing apparatus 2b obtains a decoded image by decompressing the code data b generated by the process of FIG. The process from step S601 to step S604 in FIG. 18 is the same as the process from step S601 to step S604 in FIG. 14, and the process from step S606 to step S607 in FIG. 18 is the same as the process from step S206 to step S207 in FIG. Since they are the same, the description is omitted here.

  In step S605, the inverse DC level shift and inverse color space conversion processing unit 211 performs an inverse color space conversion process based on Expression (6).

  Through the above processing, the encoded data generated by the reversible image encoding processing can be decoded, and a decoded image represented in the RGB color space can be acquired.

(Example of the configuration of a computer realizing an information processing apparatus according to an embodiment of the present invention)
FIG. 19 is an example of the configuration of a computer that executes an information processing apparatus according to an embodiment of the present invention. A computer 40 in FIG. 19 includes a CPU 41, a RAM 42, and a hard disk device 43, which are connected via a data bus. The CPU 41 controls processing executed by the computer 40 and devices connected to the computer 40. For example, the RAM 42 temporarily stores data processed by the CPU 42. The hard disk device 43 stores a large amount of data.

  The reversible image encoding process executed by the computer of FIG. 19 is according to the following procedure. First, the original image stored in the hard disk device 43 is read according to an instruction from the CPU 41 and stored in the RAM. Next, the CPU 41 executes the computer program according to the embodiment of the present invention, whereby the image code including the color space conversion method according to the embodiment of the present invention is applied to the original image stored in the RAM. Code processing is executed to generate code data. The code data is stored in the RAM 42 and then transferred to the hard disk device 43 or the like.

  On the other hand, the process of decoding the code data is performed in the reverse procedure of the image encoding process. That is, the code data stored in the hard disk device 43 is read according to an instruction from the CPU 41 and stored in the RAM 42. Next, when the CPU 41 executes the computer program according to the embodiment of the present invention, the image decoding including the color space conversion method according to the embodiment of the present invention is performed on the original image stored in the RAM. Processing is executed and image data is generated. The image data is stored in the RAM 42 and then transferred to the hard disk device 43 or the like.

  The computer program according to the embodiment of the present invention may be stored in a ROM, RAM, or a recording medium that can be read by a drive device (not shown), in addition to being stored in the hard disk device 43. . In accordance with an instruction from the CPU 41, the computer program is expanded and executed in, for example, the RAM 42, whereby the color space conversion method according to the embodiment of the present invention is executed.

  Although the best mode for carrying out the invention has been described above, the present invention is not limited to the embodiment described in the best mode. Modifications can be made without departing from the spirit of the present invention.

An example of an information processing apparatus that performs conventional image encoding processing and image decoding processing. The example (the 1) of the image coding process in the information processing apparatus which concerns on one embodiment of this invention. 4 shows an example of a detailed functional configuration of a DC level shift and color space conversion processing unit 201. An example of image data input to the wavelet transform processing unit 202. An example of coefficient data in the process of two-dimensional wavelet transform processing (part 1). An example (part 2) of coefficient data in the process of the two-dimensional wavelet transform process. An example (part 3) of coefficient data in the process of the two-dimensional wavelet transform process. An example of coefficient data by two-dimensional wavelet transform processing. The figure explaining the relationship between a decomposition level and resolution. The figure explaining the process which the bit plane encoding part 204 encodes. An example of the order of codes generated by the bit-plane encoding unit 204. The example (the 2) of the image coding process in the information processing apparatus which concerns on one embodiment of this invention. The flowchart (the 1) of the example of the image encoding process by the information processing apparatus which concerns on one embodiment of this invention. The flowchart (the 1) of the example of the image decoding process by the information processing apparatus which concerns on one embodiment of this invention. The flowchart (the 2) of the example of the image encoding process by the information processing apparatus which concerns on one embodiment of this invention. The flowchart (the 2) of the example of the image decoding process by the information processing apparatus which concerns on one embodiment of this invention. FIG. 9 is a flowchart (part 3) of an example of image encoding processing by the information processing apparatus according to the embodiment of the present invention. Flow chart of an example of image decoding processing by the information processing apparatus according to the embodiment of the present invention (part 3). The figure of the example of the composition of the computer which realizes the information processor concerning one embodiment of the present invention.

Explanation of symbols

1a, 1b, 2a, 2b, 3a, 3b Information processing apparatus 101 Preprocessing means 102 Orthogonal transformation means 103 Quantization means 104 Variable length encoding means 105 Code data generation means 106 Code data analysis means 107 Variable length decoding means 108 Inverse quantum 109 Inverse orthogonal transform unit 110 Post processing unit 201 DC level shift and color space conversion processing unit 202 Wavelet transform processing unit 204 Bit plane encoding unit 205 Packet generation unit 206 Code data generation unit 211 Inverse DC level shift and reverse color space Conversion processing unit 212 Inverse wavelet conversion processing unit 214 Bit plane decoding unit 215 Packet analysis unit 216 Packet acquisition unit 21 DC level shift processing unit 22 Color space conversion unit 23 Parameter setting unit 301 Color space conversion processing unit 302 DCT / wavelet conversion processing unit 04 CAVLC / CABAC processing unit 305 NAL generation unit 306 Byte stream format generation unit 311 Inverse color space conversion processing unit 312 Inverse DCT / inverse wavelet conversion processing unit 314 CAVLD / CABAD processing unit 315 NAL analysis unit 316 NAL acquisition unit 40 Computer 41 CPU
42 RAM
43 Hard disk device

Claims (18)

A color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
Conversion from the RGB color space to the YUV color space is based on the following equation (1), where R, G, B are the components of the color vector before conversion, and Y, U, V are the components of the color vector after conversion. A color space conversion device having color space conversion means for generating the color space.

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
Conversion from the RGB color space to the YUV color space is based on the following equation (2), with the color vector components before conversion as R, G, and B, and the color vector components after conversion as Y, U, and V: A color space conversion device having color space conversion means for generating the color space.

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
Conversion from the RGB color space to the YUV color space is based on the following equation (3), where R, G, and B are the color vector components before conversion, and Y, U, and V are the color vector components after conversion. A color space conversion device having color space conversion means for generating the color space.

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
The conversion from the YUV color space to the RGB color space is based on the following equation (4), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. A color space conversion device having color space conversion means for generating the color space.

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
The conversion from the YUV color space to the RGB color space is based on the following equation (5), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. A color space conversion device having color space conversion means for generating the color space.

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion device that performs reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
The conversion from the YUV color space to the RGB color space is based on the following equation (6), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. A color space conversion device having color space conversion means for generating the color space.

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. The color space conversion device according to claim 1, wherein (α + β + γ) is a power of two.
A color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
Conversion from the RGB color space to the YUV color space is based on the following equation (1), where R, G, B are the components of the color vector before conversion, and Y, U, V are the components of the color vector after conversion. A color space conversion method including a color space conversion step generated by

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
Conversion from the RGB color space to the YUV color space is based on the following equation (2), with the color vector components before conversion as R, G, and B, and the color vector components after conversion as Y, U, and V: A color space conversion method including a color space conversion step generated by

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
Conversion from the RGB color space to the YUV color space is based on the following equation (3), where R, G, and B are the color vector components before conversion, and Y, U, and V are the color vector components after conversion. A color space conversion method including a color space conversion step generated by

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
The conversion from the YUV color space to the RGB color space is based on the following equation (4), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. A color space conversion method including a color space conversion step generated by

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
The conversion from the YUV color space to the RGB color space is based on the following equation (5), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. A color space conversion method including a color space conversion step generated by

However, α, β, and γ are integers, and α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is. A color space conversion method for performing reversible conversion between an RGB color space and a YUV color space for a color vector representing a color,
The conversion from the YUV color space to the RGB color space is based on the following equation (6), where R, G, B as the color vector components after conversion are Y, U, V as the color vector components before conversion. A color space conversion method including a color space conversion step generated by

However, α, β, and γ are integers, and α ≠ 1, β ≠ 2, γ ≠ 1, α + γ = β, α + β + γ ≠ 0,
I (x) is a function that makes an integer by rounding up the fractional part of x toward positive infinity or rounding down toward negative infinity,
It is.   The color space conversion method according to claim 7, wherein (α + β + γ) is a power of two. Color space conversion means for performing color space conversion by the color space conversion method according to any one of claims 8 to 14 for each pixel of image data;
Compression means for compressing the image data color space converted by the color space conversion means;
An information processing apparatus. Decompression means for decompressing image data compressed by the information processing apparatus according to claim 15;
Color space conversion means for performing color space conversion by the color space conversion method according to any one of claims 8 to 14, for each pixel of image data expanded by the expansion means;
An information processing apparatus.   A computer program for causing a computer to execute the color space conversion method according to any one of claims 8 to 14.   A computer-readable information recording medium in which the computer program according to claim 17 is recorded.

Images