JP2558944B2 - Noise reduction device - Google Patents

Description

TECHNICAL FIELD The present invention makes it possible to remove noise between lines in a field by utilizing the line correlation of video signals of a television, a video, a video camera and the like. The present invention relates to a noise reduction device that reduces noise without causing deterioration of resolution in the vertical direction of the screen, which is likely to occur when line correlation is used.

2. Description of the Related Art As a conventional noise reduction device, for example,
It is disclosed in Japanese Patent No. 8574.

FIG. 14 is a block diagram showing the configuration of a conventional noise reduction device.

In FIG. 14, 1 is an input terminal for a video signal, and 2 is a conversion of the input video signal from an analog signal to a digital signal.
A / D converter, 3 has two input signals, a subtraction circuit for subtracting the other input signal from one input signal, 41 is a frame memory for delaying the noise-reduced signal from the input video signal by one frame . Since the color signal of the NTSC color video signal is phase-inverted for each frame, 5 is a color signal phase shift circuit for compensating for this, and only the color signal of the video signal delayed by the frame memory 4 is phased. Invert. 6 is a subtraction circuit for subtracting the input video signal and the video signal delayed by one frame to obtain a difference signal (frame difference signal),
Reference numeral 7 is a Hadamard transform circuit for extracting a vertical component, a horizontal component, and a diagonal component from the frame difference signal. Since the output of the Hadamard transform circuit 7 has different distributions of the noise component and the motion component, reference numeral 8 is a non-linear processing circuit that utilizes this fact to extract the noise component. Reference numeral 9 is a Hadamard inverse transform circuit, and since the noise signal extracted from the non-linear processing circuit 8 is obtained by Hadamard transform, it is returned to the original time axis by Hadamard inverse transform. Reference numeral 10 is a D / A converter that converts a digital signal into an analog signal.

In the conventional noise reduction device configured as above, when the video signal is input from the input terminal 1, the A / D converter 2
Is converted into a digital signal by means of a subtractor 3, and a non-correlation component described later is subtracted from the digital signal.
Ideally, it becomes a video signal component that does not include noise components,
It is stored in the frame memory 4 and delayed for one frame. It is each pixel point of the m frame shown in FIG. Since the color signal phase of the video signal delayed by one frame is inverted from that of the one frame signal, the color signal phase shift circuit 5 performs phase compensation to invert only the phase of the color signal and then the subtractor 6, a difference signal (frame difference signal) between two video signals having the same chroma phase is obtained. Originally, when the input video signal is a still image, this frame difference signal becomes a noise component itself, and noise can be extracted without the need for the circuit described below. However, when the input video signal is a moving image, the frame difference signal is a signal in which a signal component having no frame correlation and a noise component are combined.

The method of obtaining only the noise component from the frame difference signal will be described below. The frame difference signal is divided into a reduction component, a vertical direction component, a horizontal direction component, etc. in the Hadamard conversion circuit 7. In this case, input 1
The pattern picture element is as shown in Fig. 15, and the input is Becomes The converted output of the 2 × 4 order Hadamard transformation to F _24.

F ₂₄ = H ₂ · X ₂₄ · H _{4 From the} above equation, the output F _{24 of the} Hadamard transform is obtained from the 4 × 2 order input picture element X ₂₄ . H ₂ and H ₄ are as follows.

The output from the Hadamard conversion circuit 7 becomes a conversion output of 8 components. On the other hand, since noise has no correlation, it is distributed substantially evenly over the frequencies of the eight components of the output of the Hadamard transform circuit 7. Since the noise level at the output of the Hadamard transform circuit 7 corresponds to the noise level of the input signal as is well known, only a small level noise component can be extracted from these components through the non-linear processing circuit 8. Since each component extracted by the non-linear processing circuit 8 is obtained by Hadamard transform, it is returned to the original time axis by passing through the Hadamard inverse transform circuit 9, where a parallel digital noise signal is obtained. Become. The signal obtained here is obtained by extracting only the noise component from the frame difference signal having no frame correlation, and is supplied to the subtraction circuit 3 to obtain the noise-free digital video signal by subtracting the noise component from the input video signal. Will be done. Finally, the D / A converter 10 converts the digital video signal into the original analog signal and outputs it.

In principle, a noise reduction device using such a method has a possibility of reducing noise not only in a still image in a video signal input but also in a moving image without deteriorating the input image.

Problems to be Solved by the Invention In principle, the noise reduction device described in the conventional example can reduce noise without generating an afterimage not only in a video signal input and in a still image but also a moving image.

However, since field correlation or frame correlation is used, there is a drawback that the capacity of the memory for storing information in the circuit becomes large.

In view of the above points, the present invention utilizes the line correlation of video signals to reduce the memory capacity in a circuit by using a line memory, and a noise reduction device that does not deteriorate the resolution between lines of an input signal. The purpose is to provide.

Means for Solving the Problems To achieve the above object, the present invention provides an input video signal and n
(N> 0) first subtraction means for obtaining a difference component from the output signal of the line delay means, feature extraction means for decomposing the output of the first subtraction circuit into a plurality of feature components, and at least the feature components One is a non-linear processing means for performing non-linear processing, a second subtraction means for obtaining a difference signal between the output of the non-linear processing means and the input video signal, and a delay signal for the output signal from the second subtraction means. The output of the n-line delay means or the output of the second subtraction means is used as an output signal.

Further, the present invention is obtained from first subtraction means for obtaining a difference component between the input video signal and the output signal of the first n (n> 0) line delay means, the input video signal and the two-dimensional noise extraction circuit. A second subtraction circuit for obtaining a difference signal between the output signal of one of the two output signals, the first n-line delay means for delaying the output of the second subtraction circuit, and the first A third subtraction circuit for obtaining a difference signal between the output of the n-line delay means and the output of the second n-line delay means, among the output signals of the first n-line delay means and the two-dimensional noise extraction circuit Fourth subtraction means for obtaining a difference signal from the remaining output signal, second n-line delay means for delaying the output of the fourth subtraction means, output signal of the first subtraction circuit and the third The two-dimensional noise extraction means that receives the output signal of the subtraction circuit The output of the second n-line delay means or the output of the fourth pre-calculation means is used as an output video signal, and the two-dimensional noise extraction means decomposes two input video signals into a plurality of feature components. And a non-linear processing means for performing non-linear processing on at least one signal component of the characteristic components.

With the above-described structure, the present invention obtains the in-field line difference signal from the 1-line signal of the input signal and the next 1-line signal, and divides the line-difference signal into characteristic components by a conversion means such as Hadamard conversion, Only the noise component is extracted by performing the non-linear processing. Then, by removing this substantially noise component from the input signal, noise can be reduced by using a low-capacity memory without degrading the resolution in the vertical direction.

Further, by expanding the noise reduction device utilizing the correlation between the two lines to the utilization of the correlation between the three lines, the equivalent noise reduction device described above can be provided without significantly complicating the device. It is possible to intervene twice, and the noise reduction effect can be further enhanced.

First Embodiment FIG. 1 is a block diagram of a noise reduction device according to a first embodiment of the present invention. In Fig. 1,
Those designated by the same reference numerals as in FIG. 14 are the same components and have the same functions. Reference numeral 21 is a line memory for delaying a noise-reduced signal from the input signal by one line. 22 is NTSC
This is a color signal phase shift circuit for compensating for the color difference signal of the color video signal in which the phase is inverted line by line in the field. Reference numeral 23 is a one-dimensional Hadamard transform circuit for Hadamard transforming the intra-field line difference signal.
Reference numeral 24 is a non-linear processing circuit that performs non-linear processing on the output from the one-dimensional Hadamard transform circuit 23. 25 is a non-linear processing circuit 24
Is a one-dimensional inverse Hadamard transform circuit that transforms the output of the above into the original signal component of the time axis.

The operation of the noise reduction device of this embodiment configured as described above will be described below. When a video signal is input from the input terminal 1, the A / D converter 2 converts the analog signal into a digital signal. The subtractor 3 subtracts the difference signal from the video signal of one line before from the input video signal to subtract the non-correlation component, and ideally the video signal does not include a noise component. This video signal component is stored in the line memory 21 and delayed for one line. This line is the n line shown in FIG. Since the color signal of the NTSC color video signal is phase-inverted for each line, 22 is a color signal phase shift circuit for compensating for this, and only the color signal of the n-line video signal delayed by the line memory 21. Invert the phase. The subtracter 6 obtains a difference signal (line difference signal) between the input line (n) and the previous input line (n-1 line) (line difference signal: x ₀ , x ₁ , x ₂ , x _3, ...). This line difference signal is a signal in which a signal component having no line correlation (line non-correlation component) and a noise component are combined. One-dimensional Hadamard transform circuit
Assuming that 23 handles an input signal of 1 × 4 order, the input is x ₄ = | x ₀ x ₁ x ₂ x ₃ | The converted output of the 1 × 4 order Hadamard transform is F ₄ = | F ₀ F ₁ F ₂ F ₃ | F ₄ = x ₄ · H ₄ . However, H ₄ is as follows.

This one-dimensional Hadamard-transformed signal component is divided into a horizontal reduction frequency component, a horizontal high frequency component, and the like. Since each of these components is a noise component and a line decorrelation component, only a small level noise component can be extracted from each of these components by passing through the non-linear processing circuit 24. Each signal component contains a noise component and an in-line decorrelation component. Since there are relatively many in-line decorrelation components in the low-frequency component and many noise components in the high-pass component, the input of the nonlinear processing of the low-pass component is performed. Non-linear processing is performed in which the output level with respect to the level is lowered and the output level with respect to the input level is increased with respect to the high frequency component, and substantially only the noise component is extracted from the output from the Hadamard transform circuit 23. Since each component extracted from the non-linear circuit 24 is a component on the frequency axis obtained by the one-dimensional Hadamard transform, it is returned to the original time axis by passing through the one-dimensional Hadamard inverse transform circuit 25. It is almost the noise component extracted from the line difference signal. Then, this signal is supplied to the subtraction circuit 3 and subtracted from the input signal, whereby a digital signal without noise is obtained.
Finally, the D / A converter 10 converts the digital video signal into the original analog signal and outputs it.

Here, the line correlation of the video signal will be described in detail.
Figure 3 shows the field correlation, frame correlation, and line correlation of the video signal. As determining the magnitude of each correlation,
There are two possible shifts: a shift in the signal on the time axis and a shift in distance within the television screen. First, consider the case where the input signal is a still image. In this case, since the time axis in FIG. 3 is not considered, the frame correlation becomes the highest, followed by the field correlation and the line correlation. Next, when the input signal is a moving image, the faster the moving speed of the input signal, the lower the frame correlation becomes, the field correlation becomes higher than the frame correlation, and the line correlation becomes higher than the field correlation. Since the line correlation has nothing to do with the time axis, it is the same as the correlation for a still image and the correlation for a moving image. Thus, when there is motion in the input signal, the line correlation has a higher correlation than the field correlation and the frame correlation,
Noise can be reduced by utilizing the correlation.

Here, the case of 1 × 4 order for the input pixel unit to the Hadamard conversion circuit 23 will be described. Generally, there is an input method shown in FIG. In this method, even-numbered pixels are input as the 2N-th (where N ≧ 0) Hadamard transform, and odd-numbered pixels are input as the 2N + 1-th Hadamard-transform input, and the input pixels do not overlap in the horizontal direction of the screen. Is to transform Hadamard like
Regarding the Hadamard inverse transform, the inverse transform is performed by the same input pixel method as the input Hadamard transform. Therefore, in this method, as the Hadamard inverse transform circuit, basically four circuits are required in parallel to obtain a signal of one pixel by adding and subtracting four Hadamard transform components.

On the other hand, in the present invention, the input method shown in FIG. 5 is suitable from the viewpoint of reducing the circuit scale. In this method, the 2Nth Hadamard transform takes even-numbered pixels as input, and 2N +
In the first Hadamard transform, odd-numbered input pixels are input, and one pixel is shifted in the horizontal direction to be input. Since each pixel overlaps, four Hadamard transforms are performed. In the Hadamard inverse transform of this input method, only one pixel out of the four output pixels is obtained as the output of the Hadamard inverse transform, and in the figure, only the pixel x0 is output by the Hadamard inverse transform. With such a configuration, as the Hadamard inverse transform, it is only necessary to obtain the time-axis signal for one pixel from the four Hadamard transform components, and therefore, the signal for one pixel is obtained by adding and subtracting the four Hadamard transform components. It has an advantage that only one circuit is required, and a large circuit scale reduction effect can be obtained by combining with a noise reduction circuit having a line memory configuration.

Although the noise reduction of the NTSC composite signal has been described in the present embodiment, it is clear that the present invention is not limited to the NTSC composite signal and can be constituted by a baseband signal. In this case, the color signal phase shift circuit 22 is unnecessary.

If the noise reduction device is connected in cascade, the noise reduction effect can be further enhanced.

FIG. 6 is a block diagram of a noise reduction device according to the second embodiment of the present invention. In FIG.
Those denoted by the same reference numerals as in FIG. 14 are the same components,
Do the same thing. Reference numerals 31, 32, and 33 in FIG. 6 denote input terminals for component video input signals. 34, 35, 36 are A / D that convert each input signal from analog signal to digital signal
It is a converter. 37, 38, 39 are line memories for delaying each input signal by one line. Reference numerals 40, 41, 42 are one-dimensional line extraction circuits for extracting only noise components from the line difference signal. Reference numerals 43, 44, and 45 are D / A converters that convert the output signals whose noise has been reduced from digital to analog. 46,4
Reference numerals 7 and 48 denote output terminals for outputting a noise-reduced component output signal. FIG. 7 shows a detailed configuration of the one-dimensional noise extraction circuits 40, 41, 42 shown in FIG.
A three-dimensional Hadamard transform circuit, 52 is a non-linear processing circuit for extracting only noise components, and 53 is a one-dimensional Hadamard inverse transform circuit.

The operation of the noise reduction device of this embodiment configured as described above will be described below. When video signals are input from the input terminals 31, 32, 33, the A / D converters 34, 35, 36
Each is converted from an analog signal to a digital signal. The subtractor 3 subtracts the difference signal from the video signal of one line before from the input video signal to subtract the non-correlation component, and ideally the video signal does not include a noise component.
The video signal components are stored in the line memories 37, 38 and 39, respectively, and delayed for one line. Since these three input signals are baseband signals, there is no phase shift from the next input line signal, and the subtracter 6 obtains the line difference signal of the two lines in each circuit.

The signal processing of extracting the noise component in each input signal is
The description is omitted because it is the same as that described in the first embodiment. However, the input signal processing of the Hadamard transform circuit 51, the Hadamard inverse transform circuit 53, and the input pixel unit included in the noise extraction circuits 40, 41, and 42 of the present embodiment need not be the same, and the same. Further, the respective non-linear processing circuits 52 need not be the same.

The signal from which the noise component is subtracted by each subtraction circuit 3 is
Each D / A converter 43, 44, 45 converts the digital signal into an analog signal and outputs it.

As the input pixel method of Hadamard transform, the input pixel method and the inverse transform input pixel method described in the first embodiment can be considered.

In this embodiment, the luminance signal (Y signal) and the two color difference signals (RY signal and BY signal) are used as input signals.
Signal, G signal and B signal input can be realized in the same way.

Further, if each of the component noise reduction devices is connected in cascade, the noise reduction effect can be further increased.

FIG. 8 is a block diagram of a noise reduction device according to the third embodiment of the present invention. The components in FIG. 8 designated by the same reference numerals as those in FIG. 1 are the same components and have the same functions. Reference numeral 61 is a two-dimensional noise extraction circuit, which extracts a noise component from the two-dimensional line difference signal, and its detailed configuration is shown in FIG. 62 and 63 are subtracters, which have 2 inputs
Get the line difference signal. 64 is a 1H line memory. 65
Is a color signal phase shift circuit. Reference numeral 71 in FIG. 9 is a Hadamard transform circuit for performing two-dimensional Hadamard transform. 72 is a non-linear processing circuit, which extracts only the noise component from the output from the Hadamard transform circuit. Reference numeral 73 is a Hadamard inverse transformation circuit, which returns the input signal on the frequency axis to the original signal component on the time axis.

The operation of the noise reduction device of this embodiment configured as described above will be described below. When the n line shown in FIG. 10 is input from the input terminal 1, the A / D converter 2
The analog signal is converted into a digital signal, and the subtracter 6
Then, a difference signal (line difference signal 1) from the video signal one line before ((n-1) line) is obtained.

_{x 00 = X 00 -X 10 x} 01 = X 01 -X 11 the line difference signal 1 becomes only the noise component is the same is (n-1) line and the n lines. However, except in that case, the line difference signal 1 includes an inter-line decorrelation component and a noise signal. (N-1) one line before
Since the noise component output from the two-dimensional noise extraction circuit 61 is calculated in the subtractor 3, the line signal is a signal component from which a considerable noise component has been removed. This signal is also stored in the line memory 4 and delayed for one line. Since the color signal of the NTSC color video signal is phase inverted for each line, the color signal phase shift circuit 5
Is a circuit for compensating for this, and the 1-line delay signal is a signal whose phase is compensated by the color signal phase shift circuit 5. Similarly, the subtracter 63 obtains a difference signal (line difference signal 2) between the (n-1) line signal output from the color signal phase shift circuit 5 and the signal (n-2) line signal one line before. .

x ₁₀ = X ₁₀ −X ₂₀ x ₁₁ = X ₁₁ −X _{21 With} the output of the subtracter 6 described above, two input signals of the two-dimensional Hadamard transform are obtained. As will be described later, the (n-2) line signal is ideally a signal component that does not include a noise component, and the line difference signal 2 has a smaller noise component than the previously obtained line difference signal 1.

Here, the (n-2) line is a signal obtained by again subtracting the noise component of the output of the two-dimensional noise extraction circuit 61 from the (n-1) line from which the noise component is almost removed, and the noise component is subtracted twice. Since it has been removed by (subtractor 3, subtractor 62), it is ideally a signal component that does not include a noise component. This (n-2) line signal is transferred to the 1H line memory 64.
Is delayed by one line, resulting in (n-1) one line later.
It is out of phase with the line signal. Similar to the above, the color signal shift circuit 65 compensates for this.
The output of 64 is phase-compensated by the color signal phase shift circuit 65, and one becomes an input signal of the subtractor 63 and the other becomes an output signal.

Next, the operation of the two-dimensional noise extraction circuit 61 will be described. A two-dimensional signal composed of line difference signals is Hadamard-transformed by the input pattern shown in FIG. 10 in the case of Hadamard transform of 2 × 4 order (this Hadamard transform is not necessarily 2 × 4
There is no need to limit it to the next. ). The 2 × 4 order Hadamard transform is the same as shown in the section of the prior art. Each signal component decomposed on the two-dimensional frequency axis by the Hadamard transform circuit 71 is subjected to non-linear processing by the non-linear processing circuit 72. In this case, each signal component includes a noise component and a two-dimensional non-correlation component, the line frequency non-correlation component is relatively large in the low frequency component, and the noise component is large in the high frequency component. Non-linear processing is performed to extract a noise component from the output from the Hadamard transform circuit 71. The extracted noise component is returned to the original signal on the time axis by the Hadamard inverse conversion circuit 73, and then output.

A feature of this embodiment is that since the input video signal passes through the recursive filter for removing noise twice using the 2-line memory, it is possible to efficiently reduce noise.

Here, the input pixel unit to the Hadamard transform circuit will be described by taking the 2 × 4 order as an example. First, there is the input method shown in FIG. This method is the Nth time (however: N ≧ 0)
For the Hadamard transform of, the even-numbered pixel is input, and the odd-numbered pixel is input as the input of the 2N + 1st Hadamard transform, and the Hadamard transform is performed so that the input pixels do not overlap in the vertical and horizontal directions of the screen. As for the inverse transformation, the inverse transformation is performed by the input pixel method as in the input Hadamard transformation. Next, there is the input method shown in FIG. In this method, the even-numbered pixels are input to the 2N-th Hadamard transform, the odd-numbered input pixels are input to the 2N + 1-th Hadamard transform, and each pixel is shifted vertically and horizontally. Since it overlaps with the pixel, the Hadamard transform is performed eight times. In the inverse transform of this input method, only one pixel out of the eight pixels output by the inverse transform is obtained as the output of the inverse Hadamard transform. In the figure, in order to reduce the field memory, only x ₁₀ pixels are subjected to the Hadamard inverse transform. It is output.

It should be noted that although the noise reduction device for the NTSC composite signal has been described in the present embodiment, it is not limited to the NTSC composite signal, and it is clear that the baseband signal can also be used. In this case, the color signal phase shift circuit becomes unnecessary.

Further, the noise reduction effect can be further enhanced by connecting the above noise reduction devices in cascade.

FIG. 13 is a block diagram of a noise reduction device according to the fourth embodiment of the present invention. In FIG. 13, those denoted by the same reference numerals as in FIG. 8 are the same components,
Do the same thing. 81, 82, 83, 84, 85, 86 are subtraction circuits that output subtraction signals from two input lines. 87, 88, 89 are line memories for delaying each input signal by one line. Reference numerals 90, 91, and 92 are two-dimensional noise extraction circuits that extract only noise components from the line difference signal.

 The present embodiment configured as described above will be described.

In this embodiment, the input is the same as that in the second embodiment, and the signal processing for noise reduction of each component signal is the same as that in the third embodiment, and therefore will be omitted. However, the Hadamard conversion circuit 71 and the Hadamard inverse conversion circuit included in the noise extraction circuits 90, 91, and 92 of the present embodiment.
The input signal processing and the input pixel unit of 73 need not be the same, and similarly, the respective nonlinear processing circuits 82 need not be the same.

The signals from which the noise components have been subtracted by the subtraction circuits 81, 82 and 83 are converted from digital signals to analog signals by the D / A converters 35, 36 and 37 and output.

As the input pixel method of Hadamard transform, the input pixel method and the inverse transform pixel method described in the third embodiment can be considered.

In this embodiment, the luminance signal (Y signal) and the two color difference signals (RY signal and BY signal) are used as input signals.
Signal, G signal and B signal input can be realized in the same way.

Further, if each of the component noise reduction devices is connected in cascade, the noise reduction effect can be further increased.

As described above, according to the present invention, the line difference signal in the field is decomposed into the characteristic components by using the line correlation of the video signal by using the low capacity line memory, and the characteristic extracted signal components By performing a non-linear processing on the above, and as a Hadamard inverse transform circuit, a circuit configuration for obtaining a time-axis signal for only one pixel from a plurality of Hadamard transform components is obtained. It is possible to obtain a noise reduction device that has a small resolution deterioration and a large S / N improvement effect, and its practical effect is great.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a noise eliminator according to a first embodiment of the present invention, FIGS. 2, 4, and 5 are schematic diagrams for explaining the operation of Hadamard transform, and FIG. 3 is an image. 6 is a block diagram showing the configuration of the noise reduction circuit in the second embodiment of the present invention, and FIG. 7 is an internal configuration of the one-dimensional noise extraction circuit of FIG. 8 is a block diagram showing the configuration of a noise reduction / removal device according to a third embodiment of the present invention. FIG. 9 is a block diagram showing the internal configuration of the two-dimensional noise extraction circuit of FIG. 10, FIG. 11 and FIG. 12 are schematic diagrams for explaining the operation of the Hadamard transform, FIG. 13 is a block diagram showing the configuration of the noise elimination circuit in the fourth embodiment of the present invention, and FIG. Figure 15 is a block diagram showing the configuration of the noise reduction device. It is a schematic diagram of assistance in explaining the operation of the input signal. 1 …… Signal input terminal, 2,34,35,36 …… A / D converter, 3,6,6
2,63,81 to 86 …… Subtractor, 4,21,37 to 39,64,87 to 89 …… 1H
Line memory, 5,65 …… Color signal phase shift circuit, 7 ……
Hadamard transform circuit, 8 ... Non-linear processing circuit, 9 ... Hadamard inverse transform circuit, 10,43 to 45 ... D / A converter, 11,46
~ 48 …… Terminal terminal, 22 …… Color signal phase shift circuit, 23,5
1 …… One-dimensional Hadamard transform circuit, 24,52 …… Nonlinear processing circuit, 25,53 …… One-dimensional inverse Hadamard transform circuit, 31-33
...... Component signal input terminal, 40 to 42 …… One-dimensional noise extraction circuit, 61 …… Two-dimensional noise extraction circuit, 71 …… 2
Dimensional Hadamard transform circuit, 72 ... Non-linear processing circuit, 73 ...
… Inverse Hadamard transform circuit, 90-92 …… Two-dimensional noise extraction circuit.

Claims (8)

(57) [Claims] 1. A first subtraction means for obtaining a difference component between an input video signal and an output signal of an n (n> 0) line delay means, and an output of the first subtraction means is decomposed into a plurality of characteristic components. Feature extracting means, non-linear processing means for performing non-linear processing on at least one of the feature components, inverse feature extracting means for synthesizing and converting an output signal of the non-linear processing means into an original time axis signal, and the inverse feature extracting means. Second subtraction means for obtaining a difference signal between the output of the input signal and the input video signal, and the n-line delay means for delaying the output signal from the second subtraction means. A noise reduction device using the output of the second subtraction means as an output signal,
The feature extracting means is a Hadamard transform circuit in which the 2N (N: positive integer) Hadamard transform is input with even-numbered pixels, and the 2N + 1-th Hadamard transform is input with odd-numbered input pixels in the horizontal direction. A noise reduction device having a configuration in which one pixel is shifted and input, and the inverse feature extracting means is a Hadamard inverse transform circuit and outputs only one pixel from a plurality of Hadamard transform components that are input. . 2. A noise reduction device, wherein an input video signal composed of a plurality of component signal components is passed through a noise reduction means independent for each component signal, the noise reduction means comprising the input component video signal and n (n> n).
0) first subtracting means for obtaining a difference component of the output of the line delay means from the video signal, feature extracting means for decomposing the output of the first subtracting means into a plurality of feature components, and at least one of the feature components. And a second subtracting means for obtaining a difference signal between the output from the non-linear processing means and the input component video signal.
Means for inputting the output signal from the subtraction means to the input of the n-line delay means, and the output of the n-line delay means or the output of the second subtraction means is configured as an output component video signal, respectively. In addition, the feature extracting means uses an Hadamard transform circuit to input 2N (N: positive integer) Hadamard transforms to even-numbered pixels and input 2N + 1th Hadamard transforms to odd-numbered input pixels. The configuration is such that one pixel is shifted in the horizontal direction for input, and the inverse feature extraction means is a Hadamard inverse transform circuit for outputting only one pixel from a plurality of input Hadamard transform components. Noise reduction device. 3. The noise reducing apparatus according to claim 2, wherein the input video signal composed of a plurality of component video signals comprises a luminance signal and two color difference signals. 4. The noise reduction device according to claim 2, wherein the input video signal composed of a plurality of component video signals comprises an R signal, a B signal and a G signal. 5. A first subtraction means for obtaining a difference component between an input video signal and an output signal of a first n (n> 0) line delay means, the input video signal and a two-dimensional noise extraction circuit. A second subtraction circuit that obtains a difference signal from one of the two output signals; the first n-line delay means that delays the output of the second subtraction circuit; A third subtraction circuit for obtaining a difference signal between the output of the first n-line delay means and the output of the second n-line delay means;
Second subtraction means for obtaining a difference signal between the n-line delay means and the remaining output signal of the output signals of the two-dimensional noise extraction circuit, and a second n delay means for delaying the output of the fourth subtraction means.
A line delay means; and an output signal of the first subtraction circuit and the output signal of the third subtraction circuit, the two-dimensional noise extraction means, and an output of the second n-line delay means. The output of the fourth subtraction means is used as an output video signal, and the two-dimensional noise extraction means performs a non-linear processing into a feature extraction means for decomposing two input video signals into a plurality of feature components and at least one signal component of the feature components. A noise reduction device comprising a non-linear processing means for performing. 6. A noise reduction device, wherein an input video signal composed of a plurality of component signal components is passed through a noise reduction means independent for each component signal, and the noise reduction means comprises the input component video signal and the first component video signal. first subtraction means for obtaining a difference component from the output signal of the n (n> 0) line delay means, and one of the two output signals obtained from the input component video signal and the two-dimensional noise extraction circuit Second subtraction circuit for obtaining the difference signal with respect to the output signal, the first n-line delay means for delaying the output of the second subtraction circuit, the output of the first n-line delay means and the A third subtraction circuit for obtaining a difference signal from the output of the second n-line delay means and a difference signal between the first n-line delay means and the remaining output signals of the output signals of the two-dimensional noise extraction circuit. Fourth subtraction means, second n-line delay means for delaying the output of the fourth subtraction means, the output of the first subtraction circuit and the output signal of the third subtraction circuit as input signals. And the output of the second n-line delay means or the fourth
The output of the subtraction means of is the output component video signal,
The two-dimensional noise extraction means is a noise reduction device comprising a feature extraction means for decomposing two input video signals into a plurality of feature components and a non-linear processing means for performing non-linear processing on at least one signal component of the feature components. 7. The noise reducing apparatus according to claim 6, wherein the input video signal composed of a plurality of component video signals comprises a luminance signal and two color difference signals. 8. The noise reducing apparatus according to claim 6, wherein the input video signal composed of a plurality of component video signals comprises an R signal, a B signal and a G signal.