Image color conversion method and data storage method thereof
Patent Information
- Application Number
- CN202111596747.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-24
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-12-24
AI Technical Summary
然而,受限于cpu的访问规则,所述8个三维数值项至少需要4次读操作才能获取,这对处理的实时性存在一定的影响
[0027] The invention provides an image color conversion and its acceleration method, which utilizes the correlation between three-dimensional array items on the image conversion table to perform group and block storage, thereby enabling trilinear interpolation to be performed through one read operation. All eight required three-dimensional array items are read out to achieve hardware acceleration. Compared with the existing technology, the time of three read operations is saved. The image color conversion and its acceleration method provided by the present invention make real-time color conversion of videos or images acquired by cameras and other equipment possible.
Smart Images

Figure CN114266841B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to an image color conversion and an acceleration method thereof. Background Art
[0002] In the process of collecting images and videos, due to the influence of external factors such as light conditions during shooting, the obtained images often have problems such as uneven brightness, low contrast, color distortion, etc., and their quality may not meet the needs of subsequent operations. In order to improve the image quality and extract more usable information, the original image is usually enhanced before image processing. Commonly used image enhancement techniques include contrast processing, histogram correction, noise processing, edge enhancement, transformation processing, and pseudo color.
[0003] In some specific application scenarios, it is usually necessary to process the acquired images or videos in real time, which places high demands on the speed of the image enhancement algorithm. The trilinear interpolation algorithm is a method of linear interpolation on the tensor product grid of three-dimensional discrete sampling data. If the RGB values of the image pixels are stored as a three-dimensional array to form an image conversion table, the trilinear interpolation algorithm can be applied to the color transformation or enhancement of the image.
[0004] At present, the trilinear interpolation algorithm is used to perform color conversion based on software and by table lookup. For a pixel, it is necessary to first find 8 three-dimensional array items from the image conversion table and then perform trilinear interpolation on the values of the 8 three-dimensional array items to calculate and convert them into corresponding values. However, due to the access rules of the CPU, the 8 three-dimensional numerical items require at least 4 read operations to obtain, which has a certain impact on the real-time processing. Summary of the invention
[0005] In view of some or all of the problems in the prior art, the present invention provides, on one hand, a method for accelerating image color conversion, comprising:
[0006] Divide the three-dimensional array items in the image conversion table into N data groups, each data group contains L three-dimensional array items, and the difference between any two three-dimensional array items in each data group in any dimension is less than or equal to 1;
[0007] The N data groups are further divided into M data blocks and stored in different storage areas, wherein:
[0008] The different storage areas are configured to be readable simultaneously, and the data size that can be read from each storage area at a time is not less than L; and
[0009] The difference between the three-dimensional array items in each data group in each data block and the three-dimensional array items in other data groups in the data block in at least one dimension is greater than 2.
[0010] Furthermore, the formation of the image conversion table includes:
[0011] Divide the three color channels of blue, green and red into K segments respectively and mark the index values in sequence; and
[0012] A set of blue, green, and red index values are taken as a three-dimensional array item to form a three-dimensional array containing K*K*K items as an image conversion table.
[0013] Furthermore, M is a positive integer, and its value is not less than 8.
[0014] Furthermore, L is a positive integer, and its value is not greater than 8.
[0015] Furthermore, the value of K is preferably an even number.
[0016] Furthermore, the M data blocks are stored in M static random access memories respectively.
[0017] Furthermore, the M data blocks are stored in M direct read memories respectively.
[0018] Based on the above-mentioned acceleration method, the present invention provides, on the other hand, a method for image color conversion, comprising:
[0019] For each pixel, a set of values is determined based on the components of its three color channels: blue, green, and red.
[0020] According to the numerical value, obtaining, in the storage area, through a read operation, a three-dimensional array item corresponding to the numerical value and seven three-dimensional array items adjacent to the three-dimensional array item, wherein the seven adjacent three-dimensional array items refer to three-dimensional array items whose differences with the three-dimensional array item in any dimension are less than or equal to 1; and
[0021] The three-dimensional array item and its seven adjacent three-dimensional array items are subjected to trilinear interpolation to obtain converted blue, green, and red channel components.
[0022] Another aspect of the present invention provides a computer device, comprising:
[0023] a plurality of independent storage areas, the plurality of independent storage areas being configured to be read simultaneously;
[0024] a memory for storing a computer program; and
[0025] A processor is used to execute the computer program to implement the steps of the image color conversion method as described above.
[0026] Another aspect of the present invention provides a computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement the steps of the image color conversion method as described above.
[0027] The present invention provides an image color conversion and its acceleration method, which utilizes the correlation between the three-dimensional array items on the image conversion table to perform grouping and block storage, so that when performing trilinear interpolation, the required eight three-dimensional array items can be read out through a single read operation to achieve hardware acceleration. Compared with the existing technology, the time of three read operations is saved. The image color conversion and its acceleration method provided by the present invention make it possible to perform real-time color conversion of videos or images obtained by devices such as cameras. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] To further illustrate the above and other advantages and features of various embodiments of the present invention, a more specific description of various embodiments of the present invention will be presented with reference to the accompanying drawings. It will be understood that these drawings only depict typical embodiments of the present invention and are therefore not to be considered as limiting the scope thereof. In the accompanying drawings, for clarity, identical or corresponding parts will be represented by identical or similar reference numerals.
[0029] Figure 1 A schematic flow chart showing a method for accelerating image color conversion according to an embodiment of the present invention;
[0030] Figure 2 A schematic diagram showing a three-dimensional perspective of an image conversion table according to an embodiment of the present invention; and
[0031] Figure 3 A schematic flow chart showing a method for image color conversion according to an embodiment of the present invention. DETAILED DESCRIPTION
[0032] In the following description, the present invention is described with reference to various embodiments. However, those skilled in the art will recognize that the various embodiments can be implemented without one or more specific details or with other replacement and / or additional methods. In other cases, well-known operations are not shown or described in detail to avoid blurring the inventive point of the present invention. Similarly, for the purpose of explanation, specific configurations are described to provide a comprehensive understanding of the embodiments of the present invention. However, the present invention is not limited to these specific details. In addition, it should be understood that the various embodiments shown in the drawings are illustrative representations and are not necessarily drawn in correct proportions.
[0033] In this specification, reference to "one embodiment" or "the embodiment" means that a particular feature, structure, or characteristic described in conjunction with the embodiment is included in at least one embodiment of the present invention. The phrase "in one embodiment" appearing in various places in this specification does not necessarily all refer to the same embodiment.
[0034] In addition, the numbering of the steps of the methods of the present invention does not limit the execution order of the method steps. Unless otherwise specified, the method steps can be executed in different orders.
[0035] In the present invention, the term "adjacent three-dimensional array items" means that the difference between two three-dimensional array items in any dimension is less than or equal to 1, that is, there are no other three-dimensional array items between the two three-dimensional array items, that is, there are no intermediate data elements; and the term "non-adjacent data groups" means that the difference between the three-dimensional array items in two data groups in at least one dimension is greater than 2.
[0036] The image conversion table usually includes K*K*K three-dimensional array items, which are formed by dividing the component values (RGB values) of the three color channels of the pixel point, blue, green, and red, into K segments, and then marking the index values in sequence, and taking a set of blue, green, and red index values as the index of a three-dimensional array, thereby obtaining a three-dimensional array of K*K*K items to form an image conversion table. Each three-dimensional number group is recorded as LUT[idx_b][idx_g][idx_r], where idx_b, idx_g, and idx_r represent the blue, green, and red index values, respectively. Taking 8-bit RGB values as an example, for an 8*8*8 image conversion table, the component values of the three color channels of blue, green, and red are divided into 8 segments, where idx_b=0 indicates that the blue channel component value is 0-31, idx_g=0 indicates that the green channel component value is 0-31, and idx_r=0 indicates that the red channel component value is 0-31. Then the three-dimensional array item LUT[0][0][0] represents the set of all pixel points whose blue, green, and red channel component values are all in the range of 0-31. Similarly, the three-dimensional array item LUT[7][7][7] represents the set of all pixel points whose blue, green, and red channel component values are all in the range of 224-255. For RGB values of other digits, and / or image conversion tables of different sizes, the blue, green, and red channel component values corresponding to each three-dimensional array item contained therein are slightly different, which can be inferred based on common knowledge in the art. In the prior art, the image conversion table is usually stored in memory in sequence, such as
[0037] As shown in Table 1:
[0038] LUT[0][0][0] LUT[0][0][1] LUT[0][0][2] LUT[0][0][3] LUT[0][0][4] LUT[0][0][5] LUT[0][0][6] LUT[0][0][7] LUT[0][1][0] LUT[0][1][1] LUT[0][1][2] LUT[0][1][3] LUT[0][1][4] LUT[0][1][5] LUT[0][1][6] LUT[0][1][7] LUT[0][2][0] LUT[0][2][1] LUT[0][2][2] LUT[0][2][3] LUT[0][2][4] LUT[0][2][5] LUT[0][2][6] LUT[0][2][7] LUT[0][3][0] LUT[0][3][1] LUT[0][3][2] LUT[0][3][3] LUT[0][3][4] LUT[0][3][5] LUT[0][3][6] LUT[0][3][7] LUT[0][4][0] LUT[0][4][1] LUT[0][4][2] LUT[0][4][3] LUT[0][4][4] LUT[0][4][5] LUT[0][4][6] LUT[0][4][7] LUT[0][5][0] LUT[0][5][1] LUT[0][5][2] LUT[0][5][3] LUT[0][5][4] LUT[0][5][5] LUT[0][5][6] LUT[0][5][7] LUT[0][6][0] LUT[0][6][1] LUT[0][6][2] LUT[0][6][3] LUT[0][6][4] LUT[0][6][5] LUT[0][6][6] LUT[0][6][7] LUT[0][7][0] LUT[0][7][1] LUT[0][7][2] LUT[0][7][3] LUT[0][7][4] LUT[0][7][5] LUT[0][7][6] LUT[0][7][7]
[0039] LUT[1][0][0] LUT[1][0][1] LUT[1][0][2] LUT[1][0][3] LUT[1][0][4] LUT[1][0][5] LUT[1][0][6] LUT[1][0][7] LUT[1][1][0] LUT[1][1][1] LUT[1][1][2] LUT[1][1][3] LUT[1][1][4] LUT[1][1][5] LUT[1][1][6] LUT[1][1][7] LUT[1][2][0] LUT[1][2][1] LUT[1][2][2] LUT[1][2][3] LUT[1][2][4] LUT[1][2][5] LUT[1][2][6] LUT[1][2][7] LUT[1][3][0] LUT[1][3][1] LUT[1][3][2] LUT[1][3][3] LUT[1][3][4] LUTs[1][3][5] LUT[1][3][6] LUT[1][3][7] LUT[1][4][0] LUT[1][4][1] LUT[1][4][2] LUT[1][4][3] LUT[1][4][4] LUT[1][4][5] LUT[1][4][6] LUT[1][4][7] LUT[1][5][0] LUT[1][5][1] LUT[1][5][2] LUT[1][5][3] LUT[1][5][4] LUT[1][5][5] LUT[1][5][6] LUT[1][5][7] LUT[1][6][0] LUT[1][6][1] LUT[1][6][2] LUT[1][6][3] LUT[1][6][4] LUT[1][6][5] LUT[1][6][6] LUT[1][6][7] LUT[1][7][0] LUT[1][7][1] LUT[1][7][2] LUT[1][7][3] LUT[1][7][4] LUT[1][7][5] LUT[1][7][6] LUT[1][7][7]
[0040] …
[0041] LUT[7][0][0] LUT[7][0][1] LUT[7][0][2] LUT[7][0][3] LUT[7][0][4] LUT[7][0][5] LUT[7][0][6] LUT[7][0][7] LUT[7][1][0] LUT[7][1][1] LUT[7][1][2] LUT[7][1][3] LUT[7][1][4] LUT[7][1][5] LUT[7][1][6] LUT[7][1][7] LUT[7][2][0] LUT[7][2][1] LUT[7][2][2] LUT[7][2][3] LUT[7][2][4] LUT[7][2][5] LUT[7][2][6] LUT[7][2][7] LUT[7][3][0] LUT[7][3][1] LUT[7][3][2] LUT[7][3][3] LUT[7][3][4] LUT[7][3][5] LUT[7][3][6] LUT[7][3][7] LUT[7][4][0] LUT[7][4][1] LUT[7][4][2] LUT[7][4][3] LUT[7][4][4] LUT[7][4][5] LUT[7][4][6] LUT[7][4][7] LUT[7][5][0] LUT[7][5][1] LUT[7][5][2] LUT[7][5][3] LUT[7][5][4] LUT[7][5][5] LUT[7][5][6] LUT[7][5][7] LUT[7][6][0] LUT[7][6][1] LUT[7][6][2] LUT[7][6][3] LUT[7][6][4] LUT[7][6][5] LUT[7][6][6] LUT[7][6][7] LUT[7][7][0] LUT[7][7][1] LUT[7][7][2] LUT[7][7][3] LUT[7][7][4] LUT[7][7][5] LUT[7][7][6] LUT[7][7][7]
[0042] Table 1
[0043] When the trilinear interpolation algorithm is needed to perform color transformation, for each pixel, first determine a set of index values index(a,b,c), where a is the component index idx_b of the blue channel of the pixel, b is the component index idx_g of the green channel of the pixel, and c is the component index idx_r of the red channel of the pixel, then access the image conversion table through the CPU, read out the table values corresponding to the three-dimensional array item LUT[a][b][c] corresponding to the index value and its adjacent seven three-dimensional array items LUT[a][b][c+1], LUT[a][b+1][c], LUT[a][b+1][c+1], LUT[a+1][b][c], LUT[a+1][b][c+1], LUT[a+1][b][c+1], LUT[a+1][b+1][c], and LUT[a+1][b+1][c+1], and finally use the values in these eight three-dimensional items to perform the trilinear interpolation algorithm to obtain the converted component values of the three color channels of blue, green, and red. Because the CPU is more friendly to sequential memory access, under this storage method, the following four read operations are required to obtain the 8 table lookup values:
[0044] First read operation: read LUT[a][b][c], LUT[a][b][c+1];
[0045] Second read operation: read LUT[a][b+1][c], LUT[a][b+1][c+1];
[0046] The third read operation: read LUT[a+1][b][c], LUT[a+1][b][c+1]; and
[0047] Fourth read operation: read LUT[a+1][b+1][c], LUT[a+1][b+1][c+1].
[0048] The above operation is a pure software operation. When there are many image or video pixels to be processed, the processing speed of the method is slow. In order to speed up the processing speed to meet the real-time requirements, it is possible to consider reducing the number of read operations and obtain the required 8 table lookup values with the least number of read operations. Based on this, the inventor provides an accelerated method for image color conversion, which improves the storage method of the image conversion table through hardware, so that the 8 values required by the trilinear interpolation algorithm can be read out in one read operation.
[0049] The scheme of the present invention is further described below in conjunction with the accompanying drawings of the embodiments.
[0050] Figure 1 FIG. 1 is a flow chart showing a method for accelerating image color conversion according to an embodiment of the present invention. Figure 1As shown, a method for accelerating image color conversion includes:
[0051] First, in step 101, the three-dimensional array items in the image conversion table are divided into N data groups. The three-dimensional array items in the image conversion table with adjacent relationships are taken as a data group, and then the image conversion table is divided into N data groups, wherein the number of three-dimensional array items contained in each data group is L. The image conversion table includes multiple three-dimensional array items, and the three values in each three-dimensional array item are blue, green, and red index values, respectively. In one embodiment of the present invention, the image conversion table includes K*K*K three-dimensional array items, that is, the image conversion table divides the three color channels of blue, green, and red into K segments, and starts from 0 to K-1, which are the index values of each color channel segment. In one embodiment of the present invention, the value of K is 8, but it should be understood that in other embodiments of the present invention, the value of K can also be other values, but preferably an even number. At the same time, the three color channels of blue, green, and red can also be divided into different numbers of segments, and / or the number of points contained in each segment can be the same or different, but preferably the three color channels of blue, green, and red are evenly divided into the same number of segments. Figure 2 Taking an 8*8*8 image conversion table as an example, a three-dimensional perspective diagram of an image conversion table of an embodiment of the present invention is shown. As shown in the figure, the blue, green, and red index values idx_b, idx_g, and idx_r are used as coordinate axes to form a coordinate system. The image conversion table can be regarded as a plurality of points discretely distributed in the coordinate system. The points with adjacent relationships can be understood as two points whose difference in any dimension is less than or equal to 1. For any point in the figure, there are a total of 7 points that satisfy the adjacent relationship. Therefore, in one embodiment of the invention, the value of L should not be greater than 8, and is preferably 8. That is to say, each data group contains at most 8 three-dimensional array items. For ease of understanding, in Figure 2 In , a cube is used to represent a data group, and a vertex of the cube is a point corresponding to a three-dimensional array item; and
[0052] Next, in step 102, the N data groups are divided into M data blocks. The N data groups divided in step 101 are further divided into M data blocks and stored in different storage areas, wherein the storage areas satisfy:
[0053] The different storage areas are configured to be readable simultaneously; and
[0054] The data size that can be read at a time in each storage area is not less than L. Based on this, in one embodiment of the present invention, a method of adding a hardware storage space is adopted to meet this requirement. The hardware storage space can be, for example, a static random access memory SRAM or a direct read memory DDR, that is, the M data blocks are respectively stored in M static random access memories or direct read memories;
[0055] The M data blocks satisfy the following conditions: the data groups in each data block are not adjacent, that is, the difference between the three-dimensional array items in each data group in each data block and the three-dimensional array items in other data groups in the data block in at least one dimension is greater than 2. Figure 2 The embodiment shown in the figure is used as an example for specific description. According to the three-dimensional perspective shown in the figure, it can be clearly seen that the two data groups whose difference in at least one dimension of the three-dimensional array items is greater than 2 are the two non-adjacent cubes shown in the figure. Therefore, the condition satisfied by the M data blocks can be understood as that each data block does not include two adjacent cubes. According to the three-dimensional geometric relationship, the data group needs to be divided into at least 8 groups to meet the above conditions. Therefore, in the embodiment of the present invention, the value of M is not less than 8. Figure 2 In the illustrated embodiment, the data groups, i.e., cubes, are divided into 8 groups, as indicated by reference numerals 1 to 8 in the figure.
[0056] At this point, the storage of the three-dimensional array items in the image conversion table is completed according to the position correlation, which makes it possible to obtain the eight values required by the trilinear interpolation algorithm with only one read operation. Specifically, taking each data group containing 8 three-dimensional array items as an example, the eight values required by the trilinear interpolation algorithm may exist in the following four situations:
[0057] The eight values are in the same data group. In this case, only one read operation is required on one storage area to obtain the eight values at one time.
[0058] The eight values are distributed in two data groups. According to the data group division method as described above, the two data groups must be divided into different data blocks. Therefore, at this time, it is necessary to perform a read operation on the two storage areas at the same time to obtain the eight values.
[0059] The eight values are distributed in four data groups. According to the data group division method as described above, the four data groups must be divided into different data blocks. Therefore, at this time, it is necessary to perform a read operation on the four storage areas at the same time to obtain the eight values; and
[0060] The eight values are distributed in eight data groups. According to the data group division method as described above, the eight data groups must be divided into different data blocks. Therefore, at this time, it is necessary to perform a read operation on eight storage areas at the same time to obtain the eight values.
[0061] It can be seen that in all cases, the required 8 values can be found through one read operation, achieving hardware acceleration.
[0062] It should be understood that when the number of three-dimensional array items contained in the data group is less than 8, the number of storage areas that need to be read simultaneously may be different in various cases, but the required 8 values can still be obtained through one read operation.
[0063] Based on the acceleration method described above, Figure 3 A schematic flow chart of a method for image color conversion according to an embodiment of the present invention is shown as follows: Figure 3 As shown, a method for image color conversion includes:
[0064] First, in step 301, determine the index value. For each pixel, determine a set of index values according to the components of its three color channels: blue, green, and red. Specifically, determine in which segment of the three color channels the components of the three color channels of the pixel point fall. Taking the 8*8*8 image conversion table of 8-bit RGB values as an example, if the blue, green, and red channel component values of the pixel point are 0, 100, and 255, respectively, then the index value index (a, b, c) should be a=0, b=3, and c=7;
[0065] Next, in step 302, a plurality of three-dimensional array items are obtained. According to the index value, a plurality of required three-dimensional array items are obtained in the storage area through a read operation, specifically, the three-dimensional array item corresponding to the value and seven three-dimensional array items adjacent to the three-dimensional array item, wherein the seven adjacent three-dimensional array items are three-dimensional array items whose difference with the three-dimensional array item in any dimension is less than or equal to 1; and
[0066] Finally, in step 303, trilinear interpolation is performed to obtain converted blue, green, and red channel components by performing trilinear interpolation on the three-dimensional array item and its seven adjacent three-dimensional array items.
[0067] Another aspect of the present invention provides a computer device, comprising:
[0068] a plurality of independent storage areas, the plurality of independent storage areas being configured to be read simultaneously;
[0069] a memory for storing a computer program; and
[0070] A processor is used to execute the computer program to implement the steps of the image color conversion method as described above.
[0071] Another aspect of the present invention provides a computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement the steps of the image color conversion method as described above.
[0072] The present invention provides an image color conversion and its acceleration method, which utilizes the correlation of each point on the conversion table to perform grouping and block storage, so that when performing trilinear interpolation, all the required 8 values can be read out through a single read operation to achieve hardware acceleration. Compared with the existing technology, the time of three read operations is saved. The image color conversion and its acceleration method provided by the present invention make it possible to perform real-time color conversion of videos or images obtained by devices such as cameras.
[0073] Although various embodiments of the present invention are described above, it should be understood that they are presented as examples only and not as limitations. It is obvious to those skilled in the relevant art that various combinations, modifications and changes can be made thereto without departing from the spirit and scope of the present invention. Therefore, the breadth and scope of the present invention disclosed herein should not be limited by the exemplary embodiments disclosed above, but should only be defined according to the attached claims and their equivalents.
Claims
1. A data storage method for image color conversion, characterized in that: Includes steps: Divide the three color channels of blue, green and red into K segments respectively and mark the index values in sequence, where K is a positive integer; Take a set of blue, green, and red index values as a three-dimensional array item to form a three-dimensional array containing K*K*K items as the image conversion table; Grouping the three-dimensional array items in the image conversion table into a plurality of data groups, wherein each data group contains L three-dimensional array items, wherein L is a positive integer, and the value of L is not greater than 8, so that any two three-dimensional array items in each data group are adjacent, wherein adjacent means that the difference between the two three-dimensional array items in any dimension is less than or equal to 1, and the values of the three-dimensional array items are all integers greater than or equal to 0, and are blue, green, and red index values, respectively; Grouping the plurality of data groups into M data blocks, where M is a positive integer and the value of M is not less than 8, so that every two data groups in each data block are not adjacent, and the non-adjacent means that the difference of the three-dimensional array items in the two data groups in at least one dimension is greater than 2; and The multiple data blocks are stored in different storage areas respectively, wherein the different storage areas are configured to be able to be read simultaneously.
2. The data storage method according to claim 1, characterized in that: The value of K is an even number.
3. The data storage method according to claim 1, characterized in that: The M data blocks are respectively stored in M static random access memories or direct access memories.
4. A method for image color conversion, characterized in that: The three-dimensional array items corresponding to the image color values are stored in the data storage method according to any one of claims 1 to 3, the method comprising the steps of: Determine the blue, green, and red index values of the pixel; Determining the positions of a plurality of three-dimensional array items associated with the blue, green, and red index values; determining associated data blocks according to the positions of the plurality of three-dimensional array items; and The plurality of three-dimensional array items are read at a time from a storage area storing the data block.
5. An image color conversion method as claimed in claim 4, characterized in that: Determining the associated data blocks according to the positions of the plurality of three-dimensional array items comprises: If the plurality of three-dimensional array items are in the same data group, reading the plurality of three-dimensional array items at once from a storage area corresponding to a data block to which the data group belongs; If the plurality of three-dimensional array items are in two data groups, reading the plurality of three-dimensional array items at once from two storage areas corresponding to two data blocks to which the two data groups belong; If the plurality of three-dimensional array items are in four data groups, reading the plurality of three-dimensional array items at a time from four storage areas corresponding to four data blocks to which the four data groups belong; and If the plurality of three-dimensional array items are in eight data groups, the plurality of three-dimensional array items are read at once from eight storage areas corresponding to eight data blocks to which the eight data groups belong.
6. A computer device, characterized in that: include: a memory configured to store a computer program; A processor configured to execute the computer program to implement the steps of the image color conversion method according to any one of claims 4 to 5; as well as A plurality of independent storage areas are configured to be read simultaneously.
7. A computer-readable storage medium, characterized in that: Computer program instructions are stored thereon, and when the computer program instructions are executed by a processor, the steps of the image color conversion method as described in any one of claims 4 to 5 are implemented.
