JP2015095693A - Image processing device and image processing program - Google Patents

Abstract

Proper color conversion is performed while preventing deterioration in image quality.
An acquisition unit that acquires a plurality of analysis regions from an image, and a representative point included in each analysis region, a color conversion matrix having an eigenvector calculated based on a pixel value in the analysis region is converted, Calculated by a parameter calculation unit that calculates rotation parameters, an interpolation unit that calculates a color conversion matrix of points other than the representative points by interpolating eigenvectors, and a color conversion example and an interpolation unit calculated by the parameter calculation unit A color conversion unit that performs color conversion of an image using a color conversion matrix, and a color conversion information encoding unit that encodes color conversion information including a rotation parameter of each representative point.
[Selection] Figure 1

Description

  The present invention relates to an image processing apparatus and an image processing program for processing an image in units of blocks.

  The three components R, G, and B of the color image have a high color correlation. For this reason, in image coding, it is converted into luminance and color difference signals such as Y, Cb, and Cr, but color correlation remains after conversion. The image includes a moving image and a still image.

  Therefore, the image is subjected to color conversion by performing principal component analysis (for example, see Non-Patent Document 1) or switching between a plurality of color conversions (for example, see Non-Patent Document 2) according to the characteristics of the input image, and the image after color conversion There is a technique for encoding. The used color conversion information is sent to the encoded decoding side.

S.K.Singh, et al. "Novel adaptive color space transform and application to image compression," Signal Processing: Image Communication, vol.26, issue 10, pp.662-672. D. Marpe, et al. "An adaptive color transform approach and its application in 4: 4: 4 video coding," Proceeding of Europian Signal Processing Conference 2006.

  Here, when optimal color conversion is performed on the image, energy concentrates on the converted first component (hereinafter also referred to as C1), and the second component and the third component (hereinafter referred to as C2 and C3, respectively). The encoding efficiency is improved.

  However, the color conversion switching unit is generally a block unit. When color conversion is performed on a block-by-block basis for an image whose color changes smoothly, a block boundary appears. This image is converted to MPEG2 or AVC / H. When compression is performed by a general encoding method such as H.264, the encoding efficiency is lowered due to the loss of image continuity. Therefore, there is a problem that the image quality may be deteriorated when appropriate color conversion is performed in a general block unit.

  SUMMARY An advantage of some aspects of the invention is that it provides an image processing apparatus and an image processing program capable of performing appropriate color conversion while preventing deterioration in image quality.

  An image processing apparatus according to an aspect of the present invention includes an acquisition unit that acquires a plurality of analysis regions from an image, and eigenvectors calculated based on pixel values in the analysis regions with respect to representative points included in each analysis region. Calculated by a parameter calculation unit that converts a color conversion matrix to calculate a rotation parameter, an interpolation unit that interpolates the eigenvector to calculate a color conversion matrix of points other than the representative point, and the parameter calculation unit A color conversion unit that performs color conversion of the image using a color conversion row example and the color conversion matrix calculated by the interpolation unit, and a color conversion information code that encodes color conversion information including rotation parameters of each representative point And a conversion unit.

  The image processing apparatus includes: an energy calculating unit that calculates energy of two color components of the three color components after the color conversion; and a size of the analysis region until a predetermined condition using the energy is satisfied. The parameter calculation unit may calculate a rotation parameter based on the pixels in the analysis region reset by the region setting unit.

  The image processing apparatus may further include an image encoding unit that performs an encoding process on the color-converted image.

  An image processing apparatus according to another aspect of the present invention includes an image decoding unit that decodes an encoded image, and a color conversion matrix that is a rotation parameter of each representative point of the encoded image and has an eigenvector A color conversion information decoding unit that decodes encoded color conversion information including the converted rotation parameter, and a calculation unit that calculates a color conversion matrix for each representative point using the decoded rotation parameter; Interpolating the eigenvectors of the color conversion matrix calculated by the calculation unit to calculate a color conversion matrix of points other than the representative points, the color conversion matrix calculated by the calculation unit, and the interpolation unit A color reverse conversion unit that performs color reverse conversion of the image decoded by the image decoding unit using the calculated color conversion matrix.

  An image processing program according to another aspect of the present invention is an acquisition step of acquiring a plurality of analysis regions from an image on a computer, based on pixel values in the analysis regions for representative points included in each analysis region. A parameter calculating step of converting a color conversion matrix having the calculated eigenvector to calculate a rotation parameter; an interpolation step of interpolating the eigenvector to calculate a color conversion matrix of points other than the representative point; and the parameter calculating step. Using the calculated color conversion example and the color conversion matrix calculated in the interpolation step, a color conversion step for performing color conversion of the image, and a color conversion for encoding color conversion information including rotation parameters of each representative point An information encoding step is executed.

  An image processing program according to another aspect of the present invention includes an image decoding step for decoding an encoded image and a rotation parameter for each representative point of the encoded image, and having an eigenvector. A color conversion information decoding step for decoding the encoded color conversion information including the rotation parameter converted by the color conversion matrix; and a calculation for calculating a color conversion matrix using the decoded rotation parameter for each representative point Interpolating the eigenvector of the color conversion matrix calculated in the calculation step to calculate a color conversion matrix of points other than the representative point, the color conversion matrix calculated in the calculation step, and the interpolation step Using the calculated color conversion matrix, a color reverse conversion step for performing color reverse conversion of the image decoded in the image decoding step. , To the execution.

  According to the present invention, it is possible to perform appropriate color conversion while preventing deterioration in image quality.

1 is a block diagram illustrating an example of a configuration of an image processing device according to Embodiment 1. FIG. The figure which shows an example of the relationship between a representative point and an analysis area. The figure showing color conversion by the product of a rotation matrix. The figure which shows a rotation matrix. The figure which shows an example of parameterization to a polar coordinate system. The figure which shows an example of the normal vector which two vectors stretch. 5 is a flowchart illustrating an example of image processing in the first embodiment. FIG. 6 is a block diagram illustrating an example of a configuration of an image processing apparatus according to a second embodiment. 10 is a flowchart illustrating an example of image processing according to the second embodiment. FIG. 9 is a block diagram illustrating an example of a configuration of an image processing apparatus according to a third embodiment. 10 is a flowchart illustrating an example of image processing in the third embodiment. FIG. 10 is a block diagram illustrating an example of a configuration of an image processing apparatus according to a fourth embodiment.

  Hereinafter, each embodiment will be described in detail with reference to the accompanying drawings.

[Example 1]
First, the image processing apparatus according to the first embodiment will be described. The image processing apparatus according to the first exemplary embodiment is an apparatus that encodes an image after appropriate color conversion is performed.

<Configuration>
FIG. 1 is a block diagram illustrating an example of the configuration of the image processing apparatus 10 according to the first embodiment. An image processing apparatus 10 illustrated in FIG. 1 includes a color conversion apparatus 11 that performs color conversion, and an image encoding unit 12 that encodes the color-converted image. The color conversion device 11 alone may be referred to as an image processing device.

  The color conversion apparatus 11 includes an input unit 101, an acquisition unit 102, a parameter calculation unit 103, an interpolation unit 104, a generation unit 105, a color conversion unit 106, and a color conversion information encoding unit 107.

  The input unit 101 inputs an image. The input unit 101 outputs the input image to the acquisition unit 102 and the color conversion unit 106.

  The acquisition unit 102 extracts a plurality of representative points (pixels) from the image acquired from the input unit 101, and acquires an analysis region including the representative points. The representative point may be set in advance, or a point extracted by feature point extraction may be used as the representative point. The feature points are obtained by, for example, edge extraction.

  The acquisition unit 102 acquires an analysis region including the obtained feature point. The analysis area is, for example, a 32 × 32 area centered on the representative point. The acquisition unit 102 outputs the representative point position and the analysis region to the parameter calculation unit 103.

  The acquisition unit 102 may divide the image into a plurality of analysis regions by gradation extraction, flat portion extraction, or the like, and set an arbitrary point (for example, a center point) in the analysis region as a representative point.

  Here, the relationship between the representative point and the analysis area will be described. FIG. 2 is a diagram illustrating an example of the relationship between representative points and analysis areas. In the example illustrated in FIG. 2, the representative points are set in a grid pattern with respect to the image. The representative point is not limited to the example shown in FIG. 2, and an arbitrary point may be set by extracting feature points.

  In the example shown in FIG. 2, the analysis area ar101 is set with a predetermined area (for example, 32 × 32) centered on the representative point. Note that the analysis region is not limited to the example shown in FIG. In addition, the acquisition unit 102 may first obtain an analysis region and set an arbitrary point of the analysis region as a representative point.

  Returning to FIG. 1, the parameter calculation unit 103 converts the color conversion matrix having the eigenvector calculated based on the pixel value in the analysis region to the representative point included in each acquired analysis region, and sets the rotation parameter. calculate. That is, the parameter calculation unit 103 calculates appropriate rotation parameters θ1 to θ3 for obtaining C1 to C3 from RGB of each representative point by analyzing the analysis region.

  FIG. 3 is a diagram illustrating color conversion as a product of rotation matrices. The color conversion matrix is a 3 × 3 matrix. The color conversion matrix corresponds to the rotation of the orthogonal RGB axes. Therefore, the color conversion matrix can be regarded as a product of three rotation matrices as shown in FIG.

  FIG. 4 is a diagram illustrating a rotation matrix. The rotation matrix shown in FIG. 4 can be expressed by the following equation (1).

よって、Ｒ、Ｇ、ＢからＣ１、Ｃ２、Ｃ３への色変換は、３つの回転パラメータθ１、θ２、θ３によって決定される。

Therefore, the color conversion from R, G, B to C1, C2, C3 is determined by the three rotation parameters θ1, θ2, θ3.

  For example, the parameter calculation unit 103 calculates the eigenvector by performing principal component analysis using the pixel values in each analysis region so that the color signal correlation is most eliminated. The parameter calculation unit 103 may calculate three eigenvectors so that the energy of C3 or the energy of linear combination of C2 and C3 is minimized. The parameter calculation unit 103 converts a color conversion matrix having three eigenvectors (9 parameters) calculated by any one of the above methods into three rotation parameters and calculates. The parameter calculation unit 103 outputs the calculated eigenvector to the interpolation unit 104.

For example, the first eigenvector for the largest eigenvalue is (a _{1 (1)} , a _{1 (2)} , a _{1 (3)} ), and the R, G, B components of the input value are (R, G, B) ^T Then, C1 can be expressed as follows.

更に、２番目に大きな固有値に対する固有ベクトルを（a₂₍₂₎,a₂₍₂₎,a₂₍₃₎）とし、３番目に大きな固有値に対する固有ベクトルを（a₃₍₂₎,a₃₍₂₎,a₃₍₃₎）とし、これらの固有ベクトルを大きい順番に上から順に並べ、これを色変換行列Ａとして表せば、以下の式（２）になる。

Furthermore, the eigenvector for the second largest eigenvalue is (a _{2 (2)} , a _{2 (2)} , a _{2 (3)} ), and the eigenvector for the third largest eigenvalue is (a _{3 (2)} , a _{3 (2 )} , a _{3 (3)} ), and these eigenvectors are arranged in descending order from the top and expressed as a color conversion matrix A, the following equation (2) is obtained.

上記行列Ａは、3つの回転パラメータによって以下のようにも表せる。

The matrix A can also be expressed as follows using three rotation parameters.

ここでＳ₁はSinθ₁、Ｃ₁はCosθ₁を意味する。

Where S ₁ is Sinθ _1, C ₁ refers to Cos? _1.

  From the correspondence between the lower left components of the matrix A in Equation (2) and Equation (3), the following equation can be derived.

上記の式（４）によって、θ₂が求まる。同様にして、式（３）及び式（４）からθ₁及びθ₃を求めることができる。なお、θ₁、θ₂及びθ₃は、Sin^-1などを用いて導出されるため、複数の候補が求まる場合があるが、式（３）及び式（４）を満たすものであれば、いずれのものであってもよい。また、固有ベクトルのパラメータのサイド情報圧縮効率の向上のために、複数の候補を求めた後に、周囲のすでにθ₁〜θ₃を算出し終わったブロックのθ₁〜θ₃との差分が最も小さくなるような候補を当該ブロックのθとして決定することもできる。

Θ ₂ is obtained by the above equation (4). Similarly, θ ₁ and θ ₃ can be obtained from the equations (3) and (4). Since θ ₁ , θ _2, and θ ₃ are derived using Sin ⁻¹ or the like, a plurality of candidates may be obtained. However, as long as the equations (3) and (4) are satisfied, Any one may be used. Further, in order to improve the side information compression efficiency parameters of eigenvectors, after obtaining a plurality of candidates, the smallest difference between the theta ₁ through? ₃ block has finished calculating the already theta ₁ through? ₃ around Such a candidate can also be determined as θ of the block.

  Returning to FIG. 1, the interpolation unit 104 interpolates the eigenvector acquired from the parameter calculation unit 103 to calculate a color conversion matrix of points (interpolation target points) other than the representative points. That is, the interpolation unit 104 can calculate the weighted average of the eigenvectors of a plurality of representative points (for example, four representative points) close to the interpolation target point, thereby calculating the interpolated eigenvector of the interpolation target point.

  Alternatively, interpolation can be performed after converting the acquired three eigenvectors from the Cartesian coordinate system to the polar coordinate system as follows.

  FIG. 5 is a diagram illustrating an example of parameterization into the polar coordinate system. The interpolation unit 104 converts the eigenvectors into a polar coordinate system as shown in FIG. 5 to obtain parameters (γ, θ, φ), and interpolates these parameters using weighted averages, respectively. Find the eigenvector of a point.

  In the example shown in FIG. 5, the eigenvector orthogonality is lost, but interpolation is performed from three mutually orthogonal vectors to another three mutually orthogonal vectors, and an irregular transformation matrix is not generated. Must not.

  Further, the interpolation unit 104 may appropriately perform eigenvector interpolation by performing interpolation using Euler parameters as eigenvector interpolation. If Euler parameters are used, they can be expressed by quaternion parameters, so that rotation around a certain axis can be performed.

For example, when the interpolation unit 104 interpolates two axes, that is, when interpolation is performed using first eigenvectors of a representative point adjacent to a certain representative point, the two eigenvectors v ₁ and v ₂ are stretched. The plane normal vector is calculated from the outer product.

FIG. 6 is a diagram illustrating an example of a normal vector spanned by two vectors. As shown in FIG. 6, the interpolation unit 104 obtains a normal vector (unit vector u) between the eigenvector 1 (v ₁ ) and the eigenvector 2 (v ₂ ). The interpolation unit 104 calculates a clockwise (or counterclockwise) rotation about the axis of the obtained normal vector as a quaternion. Accordingly, the interpolation unit 104 can appropriately interpolate the eigenvector of the interpolation target point using the quaternion by using the weighted average on the Euler parameter as follows.

The relationship between eigenvector 1 (v ₁ ) and eigenvector 2 (v ₂ ) can be expressed as follows when θ is used as the quaternion q and the angle between both vectors.

ここで、i,j,kは、四元数に用いられる以下の規則を満たす数である。

Here, i, j, and k are numbers that satisfy the following rules used for quaternions.

そして、上記２つの隣り合う代表点の距離Lと、補間対象点から固有ベクトル１(v₁)に対応する代表点までの距離L₁との距離を用いて、四元数を以下のように按分することで、補間対象点の四元数ｑ_ｘを以下のように求めることができる。

Then, using the distance and the distance L of the representative points adjacent the two, the distance L ₁ to the representative point corresponding to the eigenvector 1 (v ₁₎ from the interpolation target point, apportioning the quaternion as follows By doing so, the quaternion q _x of the interpolation target point can be obtained as follows.

そして、補間対象点の固有ベクトルv_xは、四元数ｑ_ｘを用いて以下のように求めることができる。

Then, the eigenvector v _x of the interpolation target point can be obtained as follows using the quaternion q _x .

以上のようにして、補間した固有ベクトルを求めることができる。

As described above, the interpolated eigenvector can be obtained.

  The above method directly interpolated quaternions. In addition to this method, by calculating the weighted average of the logarithm of the quaternion of the eigenvectors of a plurality of representative points (for example, four representative points) close to the interpolation target point, and returning the obtained logarithmic quaternion to a real number It is also possible to calculate the interpolated eigenvector of the interpolation target point. In this case, the calculation amount can be reduced by calculating the logarithm of the quaternion of the eigenvectors of the four representative points and then performing the interpolation, as compared with the method of directly interpolating the quaternion as described above. .

  Specifically, for example, it can be calculated as follows.

First, the quaternions of the eigenvectors of the four representative points close to the interpolation target point are obtained in the same manner, and are set as q ₁ , q ₂ , q ₃ , and q ₄ respectively. The eigenvector Vx of the interpolation target point is expressed by the following equation, assuming that the interpolation coefficients by the bilinear method from the four points to the interpolation target point are α ₁ , α ₂ , α ₃ , α _4. Can be sought.

  The interpolation unit 104 outputs the color conversion matrix of each representative point and the color conversion matrix interpolated at a point other than each representative point to the generation unit 105. In addition, the interpolation unit 104 outputs the position information of each representative point and the rotation parameter of each representative point to the color conversion information encoding unit 107.

  Returning to FIG. 1, the generation unit 105 generates a color conversion map in which a color conversion matrix corresponding to each pixel position is associated with each pixel of the entire screen. The color conversion matrix is represented by, for example, three eigenvectors. The generation unit 105 outputs the generated map to the color conversion unit 106.

  The color conversion unit 106 performs color conversion on the image acquired from the input unit 101 using the color conversion map acquired from the generation unit 105. The color conversion unit 106 outputs each component after color conversion from which color correlation has been removed as much as possible to the image encoding unit 12.

  The color conversion information encoding unit 107 encodes color conversion information including a rotation parameter for each representative point. If the representative point is arbitrarily set, the color conversion information encoding unit 107 includes the position information of each representative point in the color conversion information, and if the representative point unified in advance with the decoding side is used. The position information of the representative point may not be included in the color conversion information.

  The color conversion information encoding unit 107 may use any encoding method as long as it is a lossless compression method. Since the color conversion matrix is represented by three rotation parameters and transmitted, the transmission efficiency is improved as compared with transmission by three eigenvectors having nine parameters.

  The image encoding unit 12 performs an image encoding process on each component color-converted by the color conversion device 11. The image encoding process is JPEG, MPEG-2, H.264. H.264 / AVC, H.H. A known encoding technique such as H.265 / HEVC may be used.

  The image encoded by the image encoding unit 12 and the color conversion information encoded by the color conversion information encoding unit 107 are accumulated in the storage unit or transmitted to the decoding side.

  With the above configuration, it is possible to perform appropriate color conversion while preventing deterioration in image quality.

<Operation>
Next, the operation of the image processing apparatus 10 in the first embodiment will be described. FIG. 7 is a flowchart illustrating an example of image processing in the first embodiment. In step S101 illustrated in FIG. 7, the input unit 101 inputs a captured image (a still image or a moving image).

  In step S102, the acquisition unit 102 extracts a plurality of representative points from the image, and acquires an analysis region including the representative points. Note that the acquisition unit 102 may determine the representative point after obtaining the analysis region.

  In step S103, the parameter calculation unit 103 calculates an eigenvector using the acquired pixel values in the analysis region, and calculates a rotation parameter obtained by converting a color conversion matrix having an eigenvector corresponding to the representative point. The parameter calculation unit 103 calculates a rotation parameter in each analysis region including each representative point.

  In step S104, the interpolation unit 104 interpolates a color conversion matrix other than the representative point using the eigenvector of the color conversion matrix corresponding to each representative point. For example, the interpolation unit 104 performs eigenvector interpolation processing on the eigenvectors of four representative points whose coordinates are close to the interpolation target point by performing weighted averaging according to the distances from the four representative points.

  In step S105, the generation unit 105 generates a color conversion map by associating each pixel in the image with a color conversion matrix corresponding to the pixel position. The color conversion map is, for example, a map in which each pixel is associated with an eigenvector of a color conversion matrix.

  In step S106, the color conversion unit 106 performs color conversion on each pixel in the image using a color conversion map. For example, the color conversion unit 106 specifies a color conversion matrix corresponding to each pixel of the image from the map, performs color conversion on the pixel value of each pixel using the specified color conversion matrix, and calculates a color component.

  In step S107, the image encoding unit 12 performs an image encoding process on an image having color components that have undergone color conversion. The color conversion information encoding unit 107 encodes color conversion information including position information of each representative point and a rotation parameter of each representative point. If the positions of the representative points are set to be the same on the code side and the decoding side, it is not necessary to include the position information of the representative points in the color conversion information.

  As described above, according to the first embodiment, an appropriate color conversion rotation parameter is determined at a representative point of an image, and points other than the representative point are obtained by interpolation. As a result, in the first embodiment, block conversion for color conversion is not required, and color conversion in which block boundaries do not appear can be performed, so that deterioration in image quality can be prevented.

  Also, according to the first embodiment, the color conversion matrix is represented by three rotation parameters, and these three rotation parameters are transmitted. Therefore, the transmission efficiency is higher than the transmission of eigenvectors having 9 (3 × 3) parameters. Get better.

[Example 2]
Next, an image processing apparatus according to the second embodiment will be described. In the second embodiment, processing for making the size of the analysis region appropriate is performed.

  As the analysis area is set smaller, the characteristics of the signal in the analysis area are biased, so that the energy of C2 and C3 at the representative point after color conversion tends to be smaller. However, there is a possibility that the color conversion matrix interpolated with coordinates other than the representative point is extremely far from the value when the color conversion matrix is originally optimized with the target coordinate as the representative point.

  On the other hand, as the analysis area is set larger, the bias of the signal in the analysis area decreases, and the energy of the representative points C2 and C3 after color conversion tends to increase. However, it is less likely that the color conversion matrix interpolated with coordinates other than the representative point is extremely far from the value when the color conversion matrix is originally optimized with the target image coordinates as the representative point. For example, as a specific example, when the analysis region is a large region including peripheral representative points, the influence due to interpolation tends to be small at points other than the representative points. When the analysis region is small, the characteristics of the signal for the analysis region are biased, and a color conversion matrix peculiar to the region is obtained. Therefore, the energy of C2 and C3 at the representative points after color conversion is small. There is a tendency.

  Therefore, in Example 2, after the color conversion matrix of the entire image is obtained, the color conversion matrix is applied to the image to calculate the energy amount of C2 and C3 of the entire image, and this energy amount becomes as small as possible. As described above, the size of the analysis region and the position of the representative point are determined.

<Configuration>
FIG. 8 is a block diagram illustrating an example of the configuration of the image processing apparatus 20 according to the second embodiment. An image processing apparatus 20 illustrated in FIG. 8 includes a color conversion apparatus 21 that performs color conversion, and an image encoding unit 12 that encodes an image after color conversion. Only the color conversion device 21 may be referred to as an image processing device. In the configuration shown in FIG. 8, the same components as those shown in FIG.

  The color conversion device 21 includes an input unit 101, an acquisition unit 202, a parameter calculation unit 203, an interpolation unit 204, a generation unit 205, a color conversion unit 206, a color conversion information encoding unit 107, and an energy calculation unit. 207 and an area setting unit 208.

  Unlike the first embodiment, the acquisition unit 202, the parameter calculation unit 203, the interpolation unit 204, the generation unit 205, and the color conversion unit 206 are repeatedly processed. It is the same. In this iterative process, the following process is performed in advance so that the interpolation unit 204 can perform the interpolation. That is, for example, the area setting unit 208 provisionally determines the analysis area and its representative point for the entire image. Then, it is desirable that the parameter calculation unit 203 calculates parameters for each of the provisional analysis areas so that the interpolation unit 204 can perform interpolation in this iterative process. In the iterative process, the parameters are sequentially corrected and the size of the analysis region is appropriately changed.

  A specific iterative process after the parameter calculation unit 203 calculates the parameters for each provisional analysis region is as follows.

  The energy calculation unit 207 weights and adds the energy of the color component C2 and the color component C3 in each analysis region, and calculates the total energy of the color component C2 and the color component C3 in the entire image. The energy calculation unit 203 outputs the calculated energy to the region setting unit 208.

  If the area setting unit 208 determines that the acquired energy satisfies a predetermined condition, the area setting unit 208 outputs an area size that satisfies the predetermined condition to the interpolation unit 204 and the color conversion unit 206. The predetermined condition is, for example, that the energy is minimum or minimum, or that the energy is equal to or less than a threshold value.

  The region setting unit 208 may determine whether the acquired energy is equal to or less than a threshold with respect to a predetermined condition using energy, process all sizes of the analysis region, and determine the size that provides the smallest energy. May be.

  If the acquired energy does not satisfy the predetermined condition, the area setting unit 208 sets a new analysis area by dividing the current analysis area or reducing it by a predetermined size. The region setting unit 208 notifies the acquisition unit 202 of the newly set analysis region. For example, the area setting unit 208 may divide the image into, for example, 32 × 32 as the analysis area of the initial value, and then divide the image by 16 in order of 16 × 16, 8 × 8 to reduce the analysis area.

  The acquisition unit 202 acquires a new analysis region from the region setting unit 208. The acquisition unit 202 may use the center position of the analysis region as a representative point. About the process after the parameter calculation part 203, a process is performed using a new analysis area | region.

  When the representative point is set in advance and the analysis region includes a plurality of representative points, the acquisition unit 202 may select the representative point at the center of the analysis region, for example.

  Thereby, the rotation parameter calculated using the area | region when the energy of C2 and C3 of the whole image becomes small can be calculated | required.

  The interpolation unit 204 outputs the rotation parameter of the analysis region size notified by the region setting unit 208 and information about the size of the analysis region (for example, the number of divisions) to the color conversion information encoding unit 107.

  The color conversion unit 206 performs color conversion processing using a map of a color conversion matrix corresponding to the analysis region size notified from the region setting unit 208.

  Note that the region setting unit 208 can also determine whether or not to further divide each analysis region. In this case, information about which analysis area is divided and how many times is included in the color conversion information.

<Operation>
Next, the operation of the image processing apparatus 20 in the second embodiment will be described. FIG. 9 is a flowchart illustrating an example of image processing according to the second embodiment. The processing in steps S201 to S206 shown in FIG. 9 is the same as the processing in steps S101 to S106 shown in FIG.

  In step S207, the energy calculation unit 207 adds the energy of the color component C2 and the color component C3 in each analysis region, and calculates the sum of the energy of the color component C2 and the color component C3 in the entire image.

  In step S208, the region setting unit 208 determines whether the calculated energy satisfies a predetermined condition. The predetermined condition is, for example, that the calculated energy is equal to or less than a threshold value. If the energy satisfies the predetermined condition (step S208—YES), the process proceeds to step S210. If the energy does not satisfy the predetermined condition (step S208—NO), the process proceeds to step S209.

  In step S209, the region setting unit 208 newly resets the analysis region. The area setting unit 208 resets a new analysis area, for example, by dividing the current analysis area. After step S209, the process returns to step S202, and the process is performed again in the new analysis region.

  In step S210, the image encoding unit 12 converts the analysis region in which the energy of C2 and C3 is small into an image having color components that are color-converted by a color conversion matrix having an eigenvector obtained by analysis. Then, an image encoding process is performed.

  In addition, the color conversion information encoding unit 107 encodes color conversion information including information on the size of the analysis region such that the energy between C2 and C3 is small and the rotation parameter of each representative point. The position of the representative point is obtained from the information regarding the size of the analysis area. For example, if the number of divisions is known, the size of the analysis region can be known, and the position of the representative point can be obtained by using the central point of the analysis region as the representative point.

  The color conversion information encoding unit 107 may encode the set position information of the representative point of each analysis region in the same manner as in the first embodiment.

  As described above, according to the second embodiment, in addition to the effects of the first embodiment, a color conversion matrix having an appropriate eigenvector obtained by an appropriate analysis region can be used. Therefore, encoding efficiency can be improved when image encoding is performed.

  Further, according to the second embodiment, the following can be achieved by appropriately setting the size of the analysis region and the position of the representative point using the energy amount of the color component generated by the color conversion unit. According to the second embodiment, for an image whose color changes smoothly, the representative points can be set coarsely (wider intervals) and the analysis region can be set larger. Thereby, color conversion information can be reduced and encoding efficiency improves.

  In the second embodiment, for an image with many fine textures, the representative point can be set finely (the interval is narrow), and the analysis region can be set small. This can prevent image quality deterioration.

[Example 3]
Next, an image processing apparatus according to the third embodiment will be described. The image processing apparatus according to the third embodiment generates an output image by performing a decoding process on the image that has been color-converted and encoded in the first and second embodiments and performing color inverse conversion.

<Configuration>
FIG. 10 is a block diagram illustrating an example of the configuration of the image processing apparatus 30 according to the third embodiment. An image processing device 30 illustrated in FIG. 10 includes an image decoding unit 31 that decodes an encoded image, and a color reverse conversion device 32 that performs color reverse conversion. The color inverse conversion device 32 alone may be referred to as an image processing device.

  The image decoding unit 31 performs an image decoding process corresponding to the image encoding process of the image encoding unit 12 and decodes the encoded image. The image decoding unit 31 may output a signal for synchronizing with the color conversion information corresponding to the decoded image to the color conversion information decoding unit 301.

  The color reverse conversion device 32 includes a color conversion information decoding unit 301, a calculation unit 302, an interpolation unit 303, a generation unit 304, and a color reverse conversion unit 305.

  The color conversion information decoding unit 301 decodes encoded color conversion information including, for example, rotation parameters of each representative point of an encoded image and including a rotation parameter obtained by converting a color conversion matrix having an eigenvector. To do. The color conversion information may include information regarding the position information of the representative points and the size of the analysis area. The decoding process is a decoding process corresponding to the encoding process of the color conversion information encoding unit 107.

  The color conversion information decoding unit 301 outputs the decoded information to the calculation unit 302. The color conversion information decoding unit 301 may perform the decoding process in synchronization with the synchronization signal acquired from the image decoding unit 31.

  The calculation unit 302 calculates a color conversion matrix for each representative point using the decoded rotation parameter. In addition, when the calculation unit 302 acquires the position information of the representative point and the information regarding the size of the analysis region, the calculation unit 302 obtains the position of the representative point from this information.

  For example, the calculation unit 302 specifies the size of the analysis region based on how many times it is divided, and when the image is divided by the analysis region of this size, the central point of the analysis region is used as the representative point, Can be determined.

  In addition, when the position of the representative point is determined in advance, the calculation unit 302 may use the position. The calculation unit 302 outputs a color conversion matrix corresponding to each representative point to the interpolation unit 303. The color conversion matrix at this time has three eigenvectors.

  The interpolation unit 303 interpolates the eigenvectors of the color conversion matrix calculated by the calculation unit 303, and calculates a color conversion matrix of points other than the representative points. The interpolation unit 303 performs an eigenvector interpolation process (for example, a weighted average) by using the same interpolation process as that of the first and second embodiments. The interpolation unit 303 outputs the color conversion matrix of the representative points and the color conversion matrix interpolated at points other than the representative points to the generation unit 304.

  The generation unit 304 generates a color conversion map by associating a color conversion matrix with each pixel of the image in the same manner as in the first and second embodiments. The generation unit 304 outputs the generated map to the color inverse conversion unit 305.

  The color reverse conversion unit 305 performs color reverse conversion of the image decoded by the image decoding unit 31 using the color conversion matrix calculated by the calculation unit 302 and the color conversion matrix calculated by the interpolation unit 303. The color reverse conversion unit 305 outputs the image subjected to the color reverse conversion as an output image to a storage device or a display device.

  By having the above configuration, it is possible to appropriately decode an image encoded by the image processing apparatus according to the first embodiment or the second embodiment.

<Operation>
Next, the operation of the image processing apparatus 30 in the third embodiment will be described. FIG. 11 is a flowchart illustrating an example of image processing according to the third embodiment. In step S301 illustrated in FIG. 11, the image decoding unit 31 performs an image decoding process on the input stream (encoded image data).

  In step S302, the color conversion information decoding unit 301 decodes the encoded color conversion information, and acquires position information of each representative point and rotation parameters of each representative point. Note that step S301 and step S302 are in no particular order.

  In step S303, the calculation unit 302 calculates a color conversion matrix having an eigenvector by converting the rotation parameter of the representative point for each representative point.

  In step S304, the interpolation unit 303 interpolates the color conversion matrix other than the representative point using the eigenvector of the color conversion matrix corresponding to each representative point. For example, the interpolation unit 303 performs eigenvector interpolation processing by performing a weighted average of the eigenvectors of four representative points whose coordinates are close to the interpolation target point according to the distances from the four representative points.

  In step S305, the generation unit 304 generates a color conversion map by associating each pixel in the image with a color conversion matrix corresponding to the pixel position. The color conversion map is, for example, a map in which each pixel is associated with an eigenvector of a color conversion matrix.

  In step S306, the color reverse conversion unit 305 performs color reverse conversion on each pixel in the image decoded by the image decoding unit 31 using a color conversion map. For example, the color reverse conversion unit 305 specifies a color conversion matrix corresponding to each pixel of the image from the map, and performs color reverse conversion on the color component of each pixel using the specified color conversion matrix to generate a pixel value. The color-inverted image is output as an output image to a storage device or a display device.

  As described above, according to the third embodiment, it is possible to appropriately decode the image encoded by the image processing apparatus according to the first or second embodiment.

[Example 4]
FIG. 12 is a block diagram illustrating an example of the configuration of the image processing apparatus 40 according to the fourth embodiment. An image processing apparatus 40 shown in FIG. 12 is an example of an apparatus in which the image processing described in the first to third embodiments is implemented by software.

  As illustrated in FIG. 12, the image processing apparatus 40 includes a control unit 401, a main storage unit 402, an auxiliary storage unit 403, a drive device 404, a network I / F unit 406, an input unit 407, and a display unit 408. These components are connected to each other via a bus so as to be able to transmit and receive data.

  The control unit 401 is a CPU (Central Processing Unit) that controls each device, calculates data, and processes in a computer. The control unit 401 is an arithmetic device that executes an image processing program stored in the main storage unit 402 or the auxiliary storage unit 403. The control unit 401 receives data from the input unit 407 and the storage device, calculates and processes the data, and then outputs the data to the display unit 408 and the storage device.

  Also, the control unit 401 can realize the processing described in the first to third embodiments by executing an image processing program.

  The main storage unit 402 is a ROM (Read Only Memory), a RAM (Random Access Memory), or the like. The main storage unit 402 is a storage device that stores or temporarily stores programs and data such as OS (Operating System) and application software that are basic software executed by the control unit 401.

  The auxiliary storage unit 403 is an HDD (Hard Disk Drive) or the like, and is a storage device that stores data related to application software or the like.

  The drive device 404 reads the program from the recording medium 405, for example, a flexible disk, and installs it in the storage unit.

  A predetermined program is stored in the recording medium 405, and the program stored in the recording medium 405 is installed in the image processing apparatus 40 via the drive device 404. The installed predetermined program can be executed by the image processing apparatus 40.

  The network I / F unit 406 is a peripheral having a communication function connected via a network such as a LAN (Local Area Network) or a WAN (Wide Area Network) constructed by a data transmission path such as a wired and / or wireless line. This is an interface between the device and the image processing apparatus 40.

  The input unit 407 includes a keyboard having cursor keys, numeric input, various function keys, and the like, and a mouse and a slide pad for performing key selection on the display screen of the display unit 408. The display unit 408 is configured by an LCD (Liquid Crystal Display) or the like, and performs display according to display data input from the control unit 401.

  The configuration of the image processing apparatus 10 illustrated in FIG. 1 can be realized by, for example, the control unit 401 and the main storage unit 402 as a work memory. Further, the configuration of the image processing apparatus 20 illustrated in FIG. 8 can be realized by the control unit 401 and the main storage unit 402 as a work memory, for example. The configuration of the image processing apparatus 30 shown in FIG. 10 can be realized by, for example, the control unit 401 and the main storage unit 402 as a work memory.

  The program executed by the image processing apparatus 40 has a module configuration including each unit described in the first to third embodiments. As actual hardware, when the control unit 401 reads out and executes a program from the auxiliary storage unit 403, one or more of the above-described units are loaded onto the main storage unit 402, and one or more of the respective units are main. It is generated on the storage unit 402.

  As described above, the image processing described in the first to third embodiments may be realized as a program for causing a computer to execute the image processing. The processing described in the first to third embodiments can be realized by installing this program from a server or the like and causing the computer to execute the program.

  It is also possible to record the program in the recording medium 405 and cause the computer or portable terminal to read the recording medium 405 on which the program is recorded, thereby realizing the above-described image processing.

  The recording medium 405 is a recording medium that records information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, a magneto-optical disk, etc. Various types of recording media such as a semiconductor memory for recording can be used.

  Further, each part of the image processing described in each of the above embodiments may be mounted on one or a plurality of integrated circuits.

  Each embodiment has been described in detail above. However, the present invention is not limited to the specific embodiment, and various modifications and changes other than the above-described modification are possible within the scope described in the claims. .

10, 20, 30, 40 Image processing device 11 21 Color conversion device 12 Image encoding unit 31 Image decoding unit 32 Color inverse conversion device 101 Input unit 102, 202 Acquisition unit 103, 203 Parameter calculation unit 104, 204 Interpolation unit 105, 205 generation unit 106, 206 color conversion unit 207 energy calculation unit 208 region setting unit 301 color conversion information decoding unit 302 calculation unit 303 interpolation unit 304 generation unit 305 color inverse conversion unit

Claims (6)

An acquisition unit for acquiring a plurality of analysis regions from an image;
A parameter calculation unit that converts a color conversion matrix having an eigenvector calculated based on a pixel value in the analysis region to a representative point included in each analysis region, and calculates a rotation parameter;
An interpolation unit that interpolates the eigenvector and calculates a color conversion matrix of points other than the representative points;
A color conversion unit that performs color conversion of the image using the color conversion row example calculated by the parameter calculation unit and the color conversion matrix calculated by the interpolation unit;
A color conversion information encoding unit that encodes color conversion information including a rotation parameter of each representative point;
An image processing apparatus. An energy calculation unit that calculates energy of two color components of the three color components after the color conversion;
An area setting unit that resets the size of the analysis area until a predetermined condition using the energy is satisfied,
The parameter calculation unit
The image processing apparatus according to claim 1, wherein the rotation parameter is calculated based on the pixels in the analysis region reset by the region setting unit.   The image processing apparatus according to claim 1, further comprising an image encoding unit that performs an encoding process on the color-converted image. An image decoding unit for decoding the encoded image;
A color conversion information decoding unit for decoding encoded color conversion information, including rotation parameters obtained by converting a color conversion matrix having an eigenvector, which is a rotation parameter of each representative point of the encoded image;
For each representative point, a calculation unit that calculates a color conversion matrix using the decoded rotation parameter;
An interpolation unit that interpolates eigenvectors of the color conversion matrix calculated by the calculation unit, and calculates a color conversion matrix of points other than the representative points;
A color reverse conversion unit that performs color reverse conversion of the image decoded by the image decoding unit, using the color conversion matrix calculated by the calculation unit and the color conversion matrix calculated by the interpolation unit;
An image processing apparatus. On the computer,
An acquisition step of acquiring a plurality of analysis regions from the image;
A parameter calculation step of calculating a rotation parameter by converting a color conversion matrix having an eigenvector calculated based on a pixel value in the analysis region to a representative point included in each analysis region;
An interpolation step of interpolating the eigenvector to calculate a color conversion matrix of points other than the representative points;
A color conversion step of performing color conversion of the image using the color conversion row example calculated by the parameter calculation step and the color conversion matrix calculated by the interpolation step;
A color conversion information encoding step for encoding color conversion information including a rotation parameter of each representative point;
An image processing program for executing On the computer,
An image decoding step for decoding the encoded image;
A color conversion information decoding step for decoding encoded color conversion information, including rotation parameters of each representative point of the encoded image, the rotation parameter being converted from a color conversion matrix having an eigenvector;
A calculation step for calculating a color conversion matrix using the decoded rotation parameter for each representative point,
An interpolation step of interpolating eigenvectors of the color conversion matrix calculated by the calculation step to calculate a color conversion matrix of points other than the representative points;
A color reverse conversion step for performing color reverse conversion of the image decoded by the image decoding step using the color conversion matrix calculated by the calculation step and the color conversion matrix calculated by the interpolation step;
An image processing program for executing

Images