KR0139984B1 - Digital nonlinear gain controller for a video signal - Google Patents

Abstract

The present invention relates to a digital nonlinear gain adjuster for an arbitrarily variable video signal, comprising: data storage means for storing a function value of a characteristic curve for each section and outputting a function value corresponding to an input signal, and an input signal Examining the first 1 digit from the most significant bit of, the data storage means generates a signal for selecting a function value for the start and end points of the section and a signal for selecting the number of shifters, and sets the corresponding bit to 0 in the input signal. A signal generating means for generating a signal to be generated, a calculating means for multiplying a difference of two values output from said data storing means, and a number corresponding to 0 as said bit, and a value output from said calculating means Division means for dividing according to the signal output from the signal, and the scene output from the division means Signal completion means for outputting the result of the final partial linear approximation function by adding the signal output from the data and the data storage means, and gain of the image signal by using the partial linear approximation method consisting of the interval having a power of 2 By adjusting it, a nonlinear gain adjuster can be used to freely change the characteristic curve, and a gamma compensator with a freely transformed correction curve and auto-nead, fade-in and fade-out effects that do not reduce the dynamic range can be arbitrarily variable. A digital nonlinear gain regulator for an image signal.

Description

Arbitrarily variable digital nonlinear gain regulator for video signals

1 is a graph for explaining gamma correction for correcting characteristics of an image display apparatus in a projection processing system;

2 is a graph illustrating a gamma correction process and deformation of a gamma correction curve by a table lookup method.

3 is a graph for explaining Otoni

4 is a graph for explaining fade in and fade out of a video signal;

5 is a graph for explaining a partial linear approximation function,

6 is a block diagram applying a digital nonlinear gain adjuster for an image variable that can be arbitrarily varied according to an embodiment of the present invention;

7 is a graph showing the gain control according to FIG.

The present invention relates to a digital nonlinear gain controller for an image variable that can be arbitrarily varied. More specifically, to obtain an arbitrary nonlinear characteristic curve by freely modifying a gain by using a partial linear approximation technique, gamma correction and auto-neyy of an image signal are performed. The present invention relates to a digital nonlinear gain adjuster for an arbitrarily variable video signal that enables auto knee, fade in, and fade out functions as a system.

Some functions of a system for processing video signals include gamma correction and auto-needling for smoothing the distribution of contrast, and special effect processing such as fade in and fade out to correct characteristics of an image display device.

1 is a graph for explaining gamma correction for correcting characteristics of an image display apparatus in an image processing system.

The transfer function of an image signal between cathode rays, which is an image display device in an image processing system such as a video camera, a television, and the like, represents an exponential function of gamma (γ).

That is, it has a problem that the dark signal is displayed darker with respect to the input signal, and the bright signal is brighter with respect to the input signal, as shown in FIG.

Therefore, if the input image signal has an exponential function of reciprocal gamma 1 / 마 of gamma as shown in (c) of FIG. 1, the final image output through the cathode ray tube maintains linearity as shown in (b) of FIG. The above method is gamma correction.

In the ideal system, the gamma value of the gamma correction curve is determined and fixed by the characteristics of the fluorescent material of the cathode ray tube, but in an actual image processing system, it is necessary to modify the gamma value to match the characteristics of the components of the system.

For example, since the spectral characteristics of the solid-state image pickup device in the video camera vary depending on the manufacturing company and the manufacturing process, the gamma correction curve must be modified according to the characteristics of each solid state image pickup device.

In the conventional analog method, a partial linear approximation method of changing a gamma correction curve by dividing a gamma correction curve into several sections and then adjusting an gain for each amplifier by arranging amplifiers whose gain can be varied for each section is used.

In this method, an adjustment point must be placed externally in order to vary the gain, and there is a disadvantage in that the limit is variable.

Digital processing of gamma correction requires digital processing as the image processing system becomes digital.However, when the partial linear approximation technique used in the analog method is applied in the digital system, large hardware such as a multiplier, a divider, and an adder / subtractor can be used. The table lookup method is used because of the disadvantage of using.

2 is a graph for explaining the gamma correction processing and the deformation of the gamma correction curve by the table lookup method.

As shown in FIG. 2, this method of selecting a gamma correction curve suitable for each image processing system by storing several modified gamma correction curves in advance in the storage device has the disadvantage that the gamma correction curve cannot be arbitrarily modified, and the variable gamma Since the storage device is required for the number of correction curves, the hardware size increases as the number of variable curves of the gamma correction curve increases.

3 is a graph for explaining Otoni.

If there is a component having a strong luminance signal in a part of the image, this signal causes the contrast distribution of the projection to the low light and high light components to deteriorate the image quality.

In order to prevent this, a portion having a strong luminance signal above a certain threshold lowers the gain to weaken the signal to evenly distribute the contrast.

As shown in FIG. 3, since a normal video signal has a low luminance component and an even distribution of contrast, a curve having a constant slope is selected as shown in (a) of the system.

However, if there is a strong luminance signal in the image, a characteristic curve whose slope varies with each section is selected as shown in (b) of FIG. 3.

In other words, the intensity is reduced by increasing the magnitude of the gain by lowering the gain for the strong luminance signal above the threshold and increasing the gain by increasing the gain of the other signals.

However, a multiplier and an adder of the above method are required, which leads to a disadvantage of increasing hardware.

In order to simplify the hardware configuration, the gain is set to '1' for signals below the threshold and the slope is set to '0.5' for signals above the threshold, and bit shifted to a signal having a magnitude larger than the threshold to the right. This multiplies the slope by 0.5 without using a multiplier.

However, although the size of the hardware is reduced by doing the above, there is a problem in that a dynamic range capable of expressing an image signal is reduced due to the selection of the slopes 1 and 0.5, as shown in FIG.

4 is a graph for explaining fade in and fade out of a video signal.

In processing the video signal, the fade-out method gradually decreases the size of the input video signal with time as shown in FIG. As shown in (b) of FIG. 4, the screen gradually appears as the magnitude of the input image signal is gradually increased with time.

Conventionally, by configuring a separate device including a large multiplier, an adder, and the like for the processing of the fade in and fade out as described above, there is a disadvantage in that the size of the circuit is enlarged and the circuit is complicated.

Accordingly, an object of the present invention is to solve the above problems and disadvantages, and by adjusting the gain of an image signal using a partial linear approximation method in which the magnitude of the linear section corresponds to the exponential power of 2, the variation of the characteristic curve is changed. A free non-linear gain adjuster consists of a gamma corrector with freely modified correction curves and an optional variable digital non-linear gain adjuster that handles the effects of auto knee, fade in, and fade out with no dynamic range reduction. To provide.

The configuration of the present invention for achieving the above object,

Data storage means for storing a function value of a characteristic curve for each section and outputting a function value corresponding to an input signal;

The first 1 digit is checked from the most significant bit of the input signal, and the data storage means generates a signal for selecting a function value for the start point and end point of the section and a signal for selecting the number of shifters, and the corresponding bit in the input signal. Signal generating means for generating a signal making 0 zero;

Arithmetic means for multiplying the difference between the two values output from the data storage means and the number of bits corresponding to zero and outputting the result;

Division means for dividing the value output from the calculation means according to the signal output from the signal generation means;

And a signal completion means for adding the signal output from the dividing means and the signal output from the data storage means to output the result of the final partial linear approximation function.

Hereinafter, with reference to the accompanying drawings will be described the most preferred embodiment that can be easily carried out this invention.

5 is a graph for explaining a partial linear approximation function,

6 is a block diagram of applying a digital nonlinear gain adjuster for an image variable that can be arbitrarily varied according to an embodiment of the present invention.

7 is a graph showing the gain control according to FIG.

As shown in FIG. 6, a configuration of a digital nonlinear gain regulator for an image signal which can be arbitrarily variable according to an embodiment of the present invention is

Checks the first digit from the most significant bit of the input signal, generates a signal to select the function value for the start and end points of the interval in the register, and a signal to select the number of shifters, and sets the corresponding bit to 0 in the input signal. A decoder 20 for generating a signal to make;

A register 10 for receiving a function value selection signal output from the decoder 20 and outputting a function value of an upper section and a lower section;

An arithmetic unit (30) for multiplying the difference between two values output from the register (10) by the number of bits corresponding to zero and outputting the multiplied number;

A shifter 40 which divides the value output from the operation unit 930 according to the signal output from the decoder 20;

And an adder 50 that adds a signal output from the shifter 40 and a signal output from the register 10 to output a result of the final partial linear approximation function.

The operation unit 30 may include a subtractor 31 having a function value of the lower section connected to the negative terminal and a function value of the upper section connected to the positive terminal, and an input signal and the An AND 32 having a first output terminal of the decoder 20 connected as an input, an output terminal of the subtractor 31, and an output terminal of the end 32 are connected as input terminals, and an output terminal is connected to the shifter 40. The multiplier 33 is connected to the.

The operation of the present invention made as described above is as follows.

Partial linear approximation refers to a process of simplifying the process by dividing a nonlinear characteristic curve represented by a multidimensional function into several sections and substituting each section by a linear linear function.

In FIG. 5, the characteristic curves for gamma correction, autonyi, fade in, and fade out are divided into several sections, and each section is replaced by a linear linear function using partial linear approximation.

The linear function can be obtained by substituting the value of the start point and the end point of the interval and the function value of the characteristic curve in the formula for finding the straight line connecting the two points when two points are given. The function value y 'of the characteristic curve for is given by Equation (1) below.

here,

X (n-1): the starting point of the section

X (n): end point of the interval

y (n-1): function of the characteristic curve for the beginning of the interval

y (n): function of the characteristic curve for the end point of the interval

n: exponent representing the interval

In order to implement Equation (1) in hardware, a plurality of multipliers, dividers, and adders and subtractors are required, thereby increasing the scale of the hardware.

In the present invention, since the gamma, which is the exponential multiplier of the characteristic curve, is smaller than 1, the variation of the characteristic curve becomes slower as the input signal is larger, so that the part having a large input signal has a small error even when the large part of the input signal is replaced by the first-order function of a large interval. The size of the approximation section was chosen by exponential power of 2 rather than equidistant.

If the size of the interval is an exponential power of 2, the denominator (X (n) -X (n-1) corresponding to the size of the interval in Equation (1) is calculated as shown in Equation (2) below, and the exponential power of 2 is also increased. This makes it easy to implement division as a shifter.

2 ⁿ -2 ^n-1 = 2 ^n-1 ... … … … … … … … … (2)

Therefore, the number of shifts in the above is an exponential multiplier of n (n-1).

In FIG. 5, the input signal x 'has the same relationship as in Equation (3) below.

X (n-1) = 2 ^n-1 ≦ x 'X (n) = 2 ⁿ ... … … … … … … … … (3)

In order to obtain the shift number n-1, an input signal is examined from the most significant bit (MSB) using Equation (3) to find a position where 1 is first.

Since the input signal x 'is always greater than the start point X (n-1) of the section, and the start point X (n-1) of the section is a power of (n-1) of 2, [xX] (n-1)] can be obtained by setting the (n-1) digit bit of the input signal x 'to' 0 '.

For example, if the input signal is 10 bits and its value is 262 in decimal and 0100000111 in binary, if the first digit from the most significant bit is checked, the condition is satisfied at 9 bits, so [X (n) -X (n-1)] is 0100000000 in binary 9, and [xX (n-1)] is 0000000111 in 9-bit binary.

Using the above-described principle, the required hardware size can be reduced, and arbitrary modification of the form of the linear approximation function is possible by changing the function value of the characteristic curve at the start and end points of the interval.

Referring to FIG. 6, the decoder 20 checks the number of digits in which the first 1 appears from the most significant bit of the input signal x 'and outputs it to the register 10.

The register 10 outputs a function value for a start point and an end point of a specific curve corresponding to the signal output from the decoder 20, and the function value of the start point y (n-1) is shown in (a) of FIG. ) And (b) output the function value y (n) of the end point.

The subtractor 1 of the calculator calculates the difference between the two function values y (n-1) and y (n), and the end 32 receives the signal output from the input signal x 'and the decoder 20. By logical operation, a signal such as the following expression (4) is output to (c) of FIG.

x '-X (n-1)... … … … … … … … … (4)

The multiplier 33 multiplies the signal output from the subtractor 31 and the signal output from the end 32, and a signal such as the following expression (5) is output to (d) of FIG.

(y (n) -y (n-1)) x (x'-X (n-1))... … … … … … … … … (5)

The shifter 40 exhibits the effect of division by shifting a signal such as equation (5) by the signal output from the decoder 20, and in (e) of FIG. 6, as shown in equation (6) above. The signal is output.

The adder 50 adds a function value y (n-1) of the starting point output from the register 10 to a signal such as equation (6) and outputs the most optimal signal as shown in equation (1). .

For example, in FIG. 7, when the input signal is 10 bits and the output signal is 8 bits, the input signal is divided into six sections, and for a normal image, as shown in FIG. The function values y1, y2, y3, y4, y5, and y6 of the intervals for the gamma correction curve are stored in the register 10.

However, if a strong luminance signal is present in the image and the autonyy processing is required, the function values y1 ', y2', y3 ', y4', y5 ', and y6' are stored as shown in FIG. .

By selecting the size of the section of the linear approximation function as the exponential power of 2 rather than equidistantly, the size of the section becomes 8, 16, 32, 64, 128, 156, and 512, respectively. Unlike the conventional gamma correction method, the function value By changing (y1 ~ y6, y1 '~ y6'), it is possible to generate arbitrary characteristic curves, and unlike conventional autoney, it can have 8-bit dynamic range for both normal image and image with strong luminance signal. have.

If the value of the characteristic curve stored in the register is gradually increased from 0 to time, the fade-in effect can be achieved. On the contrary, the value of the characteristic curve stored in the register can be fade-out effect gradually.

Therefore, the effect of the present invention operating as described above is to adjust the gain of the image signal by using a partial linear approximation method consisting of a section having an exponential power of 2 to configure a nonlinear gain controller with a variable characteristic curve, It is designed to provide a freely deformable gamma corrector and a digital nonlinear gain adjuster for arbitrarily variable video signals that handles the effects of autony and fade-in and fade-out without decreasing dynamic range.

Claims (6)

Arbitrarily variable digital nonlinearity for arbitrarily varying gains by using the partial linear approximation function to obtain arbitrary nonlinear characteristic curves by freely modifying the gain to achieve gamma correction, autonyi, fade in, and fade out of the video signal. Gain regulator. The digital nonlinear gain adjuster for arbitrarily variable video signal according to claim 1, wherein the partial linear approximation function sets the size of the section to an exponential power of 2. The digital nonlinear gain adjuster for an image variable, which can be arbitrarily variable, stores data of a characteristic curve of each section and outputs a function value corresponding to an input signal. Checking a first digit from the most significant bit of the input signal, generating a signal for selecting a function value for the start point and the end point of the section and a signal for selecting the number of shifters in the data storage means; Signal generating means (20) for generating a signal for making the corresponding bit zero; and arithmetic means (30) for multiplying two values output from the data storage means (10) and multiplying the number by the corresponding bit to zero; Division means (40) for dividing the value output from the calculation means (0) in accordance with the signal output from the signal generating means 20; and the division means (40) And a signal completion means 50 for outputting the result of the final partial linear approximation function by adding the signal output from the data storage means 10 and the signal outputted from the digital storage device. Gain regulator. 4. The subtracting means (31) according to claim 3, wherein the calculating means (30) comprises a lower section value output from the data storage means (0) connected to a negative terminal and an upper section value connected to a positive terminal. A second subtracting means 32 having an input signal and a first output terminal of the signal generating means 20 as an input; an output terminal of the first subtracting means 31 and the second subtracting means ( 32. A digital nonlinear gain adjuster for an arbitrarily variable video signal, characterized in that the output terminal of (32) is connected to the input terminal and the output terminal is formed of a multiplication means (33) connected to the dividing means (40). 5. The digital nonlinear gain regulator for video signal according to claim 4, wherein the second subtracting means (32) consists of logical multiplication means. 4. The digital nonlinear gain regulator of claim 3, wherein the division means (40) comprises a shifter.