JP2008097223A - Image correction and adjustment device, and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image correction and adjustment device using a correction and adjustment table in which image correction and adjustment information is recorded for properly compressing the correction and adjustment table, and for reducing the capacity required for the recording of the correction and adjustment table. <P>SOLUTION: This image correction and adjustment circuit, for correcting and adjusting and outputting an input image by a correction and adjustment table, configured of a linear section and curve section, is provided with a linear-range determining part for determining that the specific range of the correction and adjustment table is a linear range and a table generating part for generating the correction and adjustment table, by thinning the linear section with predetermined intervals. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus and method for correcting and adjusting image data.

  In recent years, digital image input devices such as color scanners and digital cameras have become widespread. When a digital image created by such a digital image input device is displayed on a display by an image processing apparatus, a phenomenon such as blackening or whitening may occur because the color reproduction range of the display is not sufficient. Further, the flat panel may not have uniform RGB characteristics, and even if the same value is input to RGB, it may not be correct white or gray, and may have a bluish or reddish color. In such a case, the image tone of the digital image, which is input data, is corrected and adjusted and output. FIG. 11 is an overall configuration diagram of an apparatus that corrects and adjusts moving image data encoded by the MPEG technology and displays it on an output device such as a flat panel. The input compressed video stream is decoded by the video decoder 1 into a YUV signal (Y is a luminance signal, U and V are color difference signals). The resolution of the decoded data is converted by the scaler 2. Data output from the scaler is output as RGB signals by the color space converter 3. The RGB signals are color-adjusted by the color adjustment unit 1101 and output to the output device 5. For correction and adjustment, a lookup table (Look Up Table; hereinafter abbreviated as LUT), which is a correction adjustment table created to refer to the relationship between input and output, is used. In addition, the video analysis unit 1102 analyzes the distribution of RGB values of the image in order to generate an LUT. The CPU 7 generates an LUT from the analysis result and the user setting value. When the LUT is recorded in the storage unit 8 such as an image processing LSI, the storage unit 8 has a greater influence on the chip area of the LSI than the logic circuit. Therefore, as the number of LUT data increases, the LSI chip is increased. The area becomes larger. As a result, the mounting area of the LSI mounted on the printed circuit board of the image processing apparatus increases, which not only causes an increase in cost, but also hinders downsizing of the image processing apparatus.

  FIG. 12 is a schematic diagram of a conventional correction adjustment circuit. In response to the input signal, the control circuit 1201 reads LUT data necessary for correction from the storage unit 12 and outputs a corrected signal.

In Patent Document 1, when it is necessary to use a plurality of different LUTs in order to obtain a certain output image, the LUTs are integrated into a single one without causing a significant increase in memory capacity and processing amount. Provides an image processing method for realizing desired image processing.
JP 2002-33922 A

  However, with the recent increase in image quality, the number of LUT layers tends to increase. For example, when the input data is RGB 3 primary color image data, assuming that the RGB of the image data is 8 bits each, there are 256 gradations per RGB primary color. Therefore, if RGB separate LUTs are stored, 256 words × 8 bits × 3 primary colors 6 A 144-bit memory capacity is required. Furthermore, when the number of layers is 12 bits, a memory capacity of 147,456 bits is required for 4,096 words × 12 bits × 3 primary colors. Furthermore, when considering luminance information and the like, the required memory capacity is further increased. Therefore, if the number of hierarchies increases, the problem of increased memory capacity cannot be solved even if LUTs are integrated into one.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an image correction adjustment apparatus and an image correction adjustment method capable of greatly reducing the memory capacity.

  In order to solve the above-described problems and achieve the object, an image correction adjustment circuit according to the present invention is an image correction adjustment circuit that corrects and outputs an input image by using a correction adjustment table configured by a straight line portion and a curve portion. A linear range determination unit that determines that the predetermined range of the correction adjustment table is a linear range, and a table generation unit that generates a correction adjustment table obtained by thinning out the linear range at a predetermined interval. And

  According to the above configuration, it is possible to thin out only the straight line correction adjustment table data for which data restoration is easy, and reduce the memory capacity required for storing the correction adjustment table without impairing the conversion accuracy of the correction adjustment table. I can do it. Further, by reducing the memory capacity, it is possible to reduce the chip area of the LSI required for memory mounting and reduce the cost.

  An embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a configuration diagram of the entire image processing apparatus according to the present embodiment. The video decoder 1 decodes the input compressed video stream. An example of the compression format is MPEG, but is not limited thereto. The scaler 2 converts the video size of the decoded data according to the output device 5 such as a flat panel. The color space converter 3 converts the YUV signal into an RGB signal and outputs it.

  The output RGB signal is input to the LUT generation unit 6 and the color adjustment calculation unit 4. The LUT generation unit 6 analyzes the RGB value distribution of the input image of one frame and generates a compressed LUT based on a compression method described later. The generated LUT is stored in the storage unit 8. Further, the color adjustment calculation unit 4 refers to the LUT stored in the storage unit 8, performs color adjustment of the RGB signal, and outputs video data to the output device 5.

  FIG. 2 is a detailed configuration diagram of the color adjustment calculation unit 4 in FIG. The color adjustment calculation unit 4 includes a range start register 100, a range information register 101, a control circuit 102, and an interpolation calculation circuit 104. In the present embodiment, the storage unit 12 is the same as the storage unit in FIG. 1, but another storage unit may be provided for LUT recording.

  The range start register 100 stores division points for dividing the LUT graph into a linear range and a curved range. The range information register 101 stores information on whether each section divided by the range start register 100 is a straight line or a curve. The storage unit 103 stores compressed LUT data for correcting and adjusting image input data.

  When RGB data is input, the control circuit 102 reads the range start register 100, the range information register 101, and the storage unit 12. The control circuit 102 calculates necessary output data from the compressed LUT data based on these read data, and further performs interpolation calculation by the interpolation calculation circuit 104 as necessary, as described later, so that after the correction adjustment. Output an image.

  FIG. 3 is a graph showing the relationship between input and output of LUT data in this embodiment. In this embodiment, it is assumed that the input data has 256 gradations of 0 to 255, and is divided into a linear range and a curved range with A = 80, B = 100, C = 160, and D = 200. In this embodiment, the linear range is thinned out every two gradations and stored. Specifically, only the LUT when the input data is an even number (0, 2, 4,...) In the linear range is stored in the storage unit 12. FIG. 4 expresses data stored in the range start register 100, and stores A, B, C, and D, which are division points between the LUT linear range and the curved range. FIG. 5 represents data stored in the range information register 101, and stores whether each range of the LUT is a linear range or a curved range. In this embodiment, the straight line is stored as 0 and the curved line is stored as 1, but the present invention is not limited to this.

FIG. 6 represents a method for determining whether or not data between two gradations can be thinned out. For example, in order to determine whether the table data T _n for the input data n can be thinned out, an interpolation operation is performed on the table data T _n−1 for the input n ₋₁ and the table data T _{n + 1} for the input n + 1. Find T _n '. Expression 1 is an expression for interpolation calculation.

'Absolute value of _{-T n | T n' T n} -T n | is determined that the curve range when larger than the threshold TH that has been predetermined, and the linear range in the case of smaller than TH judge.

FIG. 7 is a straight line range / curve range determination flow of LUT data. In step S100, the process starts from n = 0. In step S101, it is determined whether n is an odd number. In the case of an odd number, T _n ′ is calculated based on Equation 1 in S102. S103 is compared with T _n 'with a threshold value TH which calculated in, when it is determined that the curve range and the judgment value A _n = 0, when it is determined that the curve range and A _n = 1. In S104, it is determined whether or not n is 1. If it is not 1, S105 is executed. In S105, A _n obtained this time is compared with A _n−2 already obtained. _If A _n ≠ A _n−2 , n−1 is stored in the range start register (S106), and the range information register is stored. _An is stored (S107). In step S108, n is incremented. If n is 254, the input data of this embodiment is 256 gradations, and the determination process is terminated (S109).

  In this embodiment, A = 80, B = 100, C = 160, and D = 200 are stored in the range start register of FIG. 4 as described above, and the range information register of FIG. , 1 between A and B, 0 between B and C, 1 between C and D, and 0 between D and 255.

FIG. 8 is a data storage image of the storage unit 12 in this embodiment. In FIG. 8, T _n means LUT data when the input data is n. The storage capacity value for storing the LUT data is 256 words per primary color in the conventional technique. However, in this embodiment, the linear range is thinned out every two gradations, so the required capacity value is 80/2 + 20 + 60 / It can be reduced to 2 + 40 + 56/2 = 158 words. Further, the straight line determination of the LUT data with the input data = 255 cannot be determined because the input data = 256 does not exist, and is stored as it is. Therefore, the capacity value of the storage unit 12 necessary for storing the LUT in this embodiment is 158 + 1 = 159 words, which is a capacity reduction of about 40%. The reduction rate increases as the linear range increases.

  FIG. 9 is a flow for correcting and outputting input data using the LUT. An address value corresponding to the input data is calculated by an address calculation flow described later with reference to FIG. 10 (S200). When the input data is an even number or when the input data is within the range of the curve, LUT data exists in the storage unit as shown in FIG. Therefore, the table data may be read based on the calculated address value (S202), and the read data may be output as it is (S203).

  On the other hand, when the input data is an odd number and is in the linear range, the data is thinned out to reduce the storage capacity of the storage unit, and therefore does not exist in the table data of FIG. Therefore, the table data stored in the address A calculated in step S200 and the adjacent address A + 1 is read (S204). By performing linear interpolation based on the read table data (S205) and outputting an interpolation calculation result (S206), it is possible to output table data when the input data is an odd number and within the linear range.

  10A and 10B are detailed flowcharts of the address calculation step (S202, S203, and S207) in the flow of FIG. FIG. 10A is a flow for preparation for address calculation, and FIG. 10B is a flow for calculating an address from the parameter calculated by FIG. 10A and the value of the range start register of FIG. Note that n used in FIG. 10 and n in FIG. 7 are irrelevant.

In FIG. 10A, based on the value I _n of the range information register, to define a new value of the variable I _n '. For each range information register value I _n , if I _n = 0, I _n ′ = 2, and if I _n = 1, I _n ′ = 1 (S300). The process starts from n = 0 and is repeated until n = m−1 (S301). Here, m is the number of range information registers.

In FIG. 10B, the address where the data to be output is stored for the input data Din is calculated using I _n ′ obtained in FIG. 10A and the value S _n of the range start register. When data Din is input (S400), compares the magnitude relation between Din and S ₀ (S401). If Din is equal to or less than S ₀ , the address A is calculated in step S410, and the address calculation process ends.

Din is larger than S ₀ (S401), first, the address A to _{S 0 / I 0 '(S402} ). n is counted up (S403), and it is determined whether n = m-1 (S401). If n = m−1 and Din = Dmax (S406), the address A is calculated in step S409, and the address calculation process is terminated. On the other hand, if Din ≠ Dmax (S406), the address A is calculated in step S408, and the address calculation process is terminated. Note that Dmax is the maximum value of input data in the LUT, and in this embodiment, Dmax = 255.

If n ≠ m−1 and Din is _{equal to} or smaller than S _n in step S404 (S405), address A is calculated in step S408, and the address calculation process is terminated.

  On the other hand, when Din is larger than Sn (S407), the address A is calculated in step S407, and the above processing is performed again from step S403. Then, the process is repeated until n = m−1 or Din becomes Sn or less (S404, S405).

  As described above, the LUT data of the curve portion is stored as it is, and the LUT data of the straight line portion is thinned out, thereby realizing the compression of the storage capacity necessary for holding the LUT. Further, by performing the address calculation and linear interpolation calculation of FIG. 10, desired output data corresponding to the input data can be obtained from the compressed LUT data.

It is a block diagram of the whole image processing apparatus in a present Example. It is a block diagram of the correction | amendment adjustment circuit in a present Example. The LUT in a present Example is displayed as a graph. It is a memory map of the range start register in a present Example. It is a memory map of the range information register in the present embodiment. The determination of the straight line part in a present Example is graphed. It is a straight line / curve determination flow of LUT data. It is a memory map of LUT storage RAM in a present Example. It is a data correction adjustment flow in the present embodiment. It is a preparatory flow for address calculation in a present Example. It is an address calculation flow in the present embodiment. It is a block diagram of the whole image processing apparatus in a prior art. It is a block diagram of the correction adjustment circuit in a prior art.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Video decoder 2 Scaler 3 Color space converter 4 Color adjustment calculation part 5 Output device 6 LUT production | generation part 7 CPU
8 Storage Unit 100 Range Start Register 101 Range Information Register 102 Control Circuit 104 Interpolation Operation Circuit 1101 Color Adjustment Unit 1102 Video Analysis Unit 1201 Control Circuit

Claims (5)

An image correction adjustment circuit that corrects and outputs input data by a correction adjustment table configured by a straight line portion and a curve portion,
A linear range determination unit that determines that the predetermined range of the correction adjustment table is a linear range;
A table generator for generating a correction adjustment table thinned out at a predetermined interval for the linear range;
An image correction adjustment apparatus comprising: The image correction adjustment apparatus according to claim 1, further comprising: an address calculation unit that calculates an address at which the correction adjustment table is stored;
A table control unit that reads out the correction adjustment table based on a calculation result by the address calculation unit;
An interpolation calculation unit that performs an interpolation calculation on the thinned linear part of the correction adjustment table;
An image correction adjustment apparatus comprising: The table control unit
About the correction adjustment table, the absolute value of the difference between the value corresponding to the input data and the value of the output data in the correction adjustment table, which is on a straight line connecting the output data corresponding to the preceding and subsequent data adjacent to the input data, The image according to claim 1, further comprising a determination unit that determines a straight line part when a threshold value for determining a straight line part is provided and the absolute value of the difference is smaller than the threshold value. Correction adjustment device. The image correction adjustment device according to any one of claims 1 to 3, further comprising a range start register that stores a start position of a straight line portion and a curve portion of the correction adjustment table;
A range information register for storing that one range of the correction adjustment table divided by the start position is a linear portion;
An image correction adjustment apparatus comprising: An image correction adjustment method for correcting and outputting input data by a correction adjustment table configured by a straight line portion and a curve portion,
Determining that the predetermined range of the correction adjustment table is a straight line;
Generate a correction adjustment table that is thinned at a predetermined interval for the straight line part,
Calculate the address where the correction adjustment table data for the input data is stored,
Read correction adjustment table data based on the calculation result of the address,
When the correction adjustment table data is a linear portion after thinning, an interpolation calculation is performed on the linear portion.

Images