JPH08163410A - Noise reduction circuit - Google Patents

Abstract

PURPOSE: To enhance image quality of a video signal after noise reduction processing more than the image quality of a conventional circuit, regardless of an image content. CONSTITUTION: A noise reduction circuit having a unitary cyclic filter 200 is provided with a feedback ratio control means 27 evaluating a motion vector and changing a feedback ratio depending on the evaluation result. An object signal extract means extracts plural video signals having positional line correlation based on a current field video signal and a video signal of at least one- preceding field or over, and an isolated point eliminating signal selection means 29 selects any signal based on the line correlation to eliminate isolation point noise. A motion vector detection means 26 detects a motion vector, an isolation point elimination execution/non-execution control means 27 sets an isolation point elimination signal selection means to be a selecting state of a signal based on the correlation or a passing state of a current field video signal based at least on the magnitude of the motion vector.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a noise reduction circuit for television signals, and more particularly to a noise reduction circuit using motion vectors.

[0002]

2. Description of the Related Art As a conventional noise reducer (noise reduction circuit) for reducing noise of a television signal,
Known are (1) a first-order cyclic filter using a frame memory and (2) a method using an isolated point removal method.

The noise reduction circuit (1) reduces noise by utilizing the fact that a television signal has frame correlation but a noise component has no correlation. As shown in FIG. 2, this noise reduction circuit takes out the television signals of the current frame and the previous frame from before and after the one-frame delay circuit 13 and adds them to the addition circuit 11 to increase the level of the television signal component. On the other hand, the principle is to weaken the random noise component and improve the S / N ratio. The noise removal by this method is basically effective only in a still image area having a large frame correlation, and on the contrary, the image quality deteriorates in a moving image area. Therefore, the motion detection circuit 14 determines the moving image / still image area, and the multiplication circuits 10 and 12 and the multiplication coefficient generation circuit 15 add the television signals of the current frame and the previous frame in the moving image area and the still image area. In other words, the ratio changes the feedback ratio of the television signal of the previous frame to the television signal of the current frame.

Therefore, in the noise reduction circuit shown in FIG. 2, the feedback amount of the television signal of the previous frame is small in the moving image area, and the noise reduction amount is small.

In order to improve this, a noise reduction circuit using a motion vector as shown in FIG. 3 has already been proposed. In this noise reduction circuit, the motion vector detection circuit 16 detects the motion vector, and the motion correction circuit 17 uses the motion vector to correct the motion of the television signal of the previous frame and feed it back.
A signal having a large frame correlation is fed back also to the moving image area to attenuate only random noise components.

In this noise reduction circuit, the multiplication coefficient generation circuit 15 uses the difference amount between the input signal (current frame signal) and the motion correction frame signal (previous frame signal), that is, the prediction error, to add the two signals together. Therefore, the feedback amount of the previous frame signal is determined.

On the other hand, the noise reduction method of the above (2) utilizes the interlace method or line correlation of the television signal as shown in FIG. 4 to sandwich the signal of the current field and the scanning line of the current processing target. A total of three signals of the two scanning lines one field before are compared, and the signal having the second highest level among the three signals is extracted. If there is a correlation between the signals (A, B, and C in FIG. 4 show the case where there is a correlation), it is the same regardless of which one is taken out, and the second highest level signal is taken out of the current field. It is equivalent to outputting the signal as it is,
If there is no correlation (D in FIG. 4 indicates a case where there is no correlation with the two lines in the previous field), it is regarded as noise, and the signal with the second highest level is extracted, so that other correlations are large. Random noise (isolated point) that is isolated by replacing with a signal is reduced.

[0008]

However, in the conventional noise reduction circuit shown in FIG. 3, there is no limitation on the motion vector for executing the motion compensation, and the difference amount between the input signal and the motion compensation frame signal is not limited. That is, since the feedback amount is limited by the prediction error, this difference amount does not always match the motion vector detection accuracy, and even if the motion amount is captured for a signal with many noise components, the difference amount is The amount of feedback is increased, the amount of feedback is reduced, and the noise reduction effect is reduced.

Further, due to the accumulation effect of the image pickup device, the frequency band of the signal is lowered when the motion vector is increased, the difference amount between the input signal and the motion correction frame signal is also decreased, and the feedback is performed even if the motion amount is not accurately captured. The amount becomes large, and conversely, the image quality is deteriorated.

In the noise reduction circuit adopting the above method (2), if moving image detection is performed and isolated points are removed only in the still image area, noise can be reduced in the still image area, but inter-field correlation Is not improved for a small moving image area, and the moving image detection performance not only reduces noise but also causes image quality deterioration. The great deterioration of the image quality is that the signal after isolated point removal is delayed by one field. For example, as shown in FIG. 4, when the isolated point elimination filter is configured by using the signals B and C of the previous field with respect to the signal A of the current field, the assumption is different from the above assumption, but The correlation between the signal A and the preceding field signals B and C becomes small, and this filter output outputs the preceding field signal B or C when the correlation between the preceding field signals B and C is large.

[0011]

In order to solve the above-mentioned problems, the first aspect of the present invention uses a motion vector to motion-correct a previous frame video signal, and the motion-corrected previous frame video signal is converted into a current frame. A noise reduction circuit having a first-order recursive filter configuration, which feeds back a video signal at a predetermined feedback ratio to reduce a noise component contained in the current frame video signal, evaluates a motion vector, and feedbacks the motion vector according to the evaluation result. It is characterized in that a feedback ratio control means for changing the ratio is provided.

The second aspect of the present invention is a target signal extracting means for extracting a plurality of video signals having a line correlation in terms of position from a current field video signal and a video signal of at least one field before. In a noise reduction circuit for removing isolated point noise, the isolated field noise removal circuit is provided with an isolated point removal signal selecting unit that selects any one of a plurality of extracted signals based on a correlation between the signals.
Alternatively, the motion vector detecting means for detecting the motion vector for the video signal several fields before, and the isolated point removing signal selecting means based on at least the magnitude of the detected motion vector are used to select the signal operation state based on the correlation. , Or isolated point removal execution / non-execution control means for setting the current field video signal to the passing state.

[0013]

In the noise reducing circuit of the first aspect of the present invention, the motion compensation of the previous frame video signal is performed using the motion vector, and the motion compensated previous frame video signal is fed back to the current frame video signal at a predetermined feedback ratio. In order to reduce the noise component included in the current frame video signal, the feedback ratio control means evaluates the motion vector and changes the feedback ratio according to the evaluation result. As a result, the optimum feedback ratio can be set according to the still image area, the low speed moving image area, the high speed moving image area, etc. of the current frame video signal, and the image quality of the output signal can be improved compared with the conventional case.

The noise reducing circuit according to the second aspect of the present invention is
Basically, the target signal extracting means extracts a plurality of video signals having line correlation in terms of position from the current field video signal and the video signal of at least one field before, and the isolated point removing signal selecting means. However, the isolated point noise is removed by selecting any one of the plurality of extracted signals based on the correlation between these signals.
Here, the isolated point is satisfactorily removed when the correlation between a plurality of video signals extracted by the target signal extraction means is high, and the image quality is rather deteriorated by the removal operation depending on the movement of the video signal. There is. Therefore, in the second aspect of the present invention, the motion vector detecting means detects the motion vector of the video signal of the current field video signal one or several fields before, and the isolated point removal execution / non-execution control means detects at least. Based on the magnitude of the obtained motion vector, the isolated point removing signal selecting means is set to the signal selecting operation state based on the correlation or the current field video signal passing state.

The noise reduction circuit may be configured to have both the features of the first invention and the second invention.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of a noise reduction circuit according to the present invention will be described below in detail with reference to the drawings. Here, in FIG.
It is a block diagram which shows the whole structure of the noise reduction circuit of 1st Example.

In FIG. 1, a digital input signal (composite video signal) from the input terminal 21 is applied to a Y / C separation circuit 22 and separated into a luminance signal Y and a chroma signal C. The first embodiment is a noise reduction circuit for the luminance signal Y. In the first embodiment, the circuit is configured to input the composite video signal as the input signal, but the Y / C separation circuit 22 is not necessary when the input signal is the component signal.

The luminance signal Y from the Y / C separation circuit 22 is
It is provided to the noise removal control circuit 100 and the noise removal configuration after the delay circuit 28.

As will be described later, the noise reduction circuit of the first embodiment is provided with an isolated point removal filter circuit and a first-order cyclic filter as a noise removal configuration, and the noise removal control circuit 100 controls both. It generates a signal. The noise removal control circuit 100 includes a two-dimensional low-pass filter 24, a frame delay circuit 25, a motion vector detection circuit 26, and a motion vector evaluation circuit 27.

The two-dimensional low-pass filter 24 cuts out the high frequency components of the separated luminance signal Y and extracts only the low frequency components. The two-dimensional low-pass filter 24
May be omitted by using the signal when a low-frequency luminance signal is obtained as in the case of using the two-dimensional adaptive comb filter in the Y / C separation circuit 22. In the two-dimensional adaptive comb filter, a low band luminance signal is used for motion detection.

This low band luminance signal (current frame signal) is
Motion vector detection circuit 26 and motion vector evaluation circuit 2
7 is directly applied to the frame 7, and is delayed by one frame via the one-frame delay circuit 25 and then applied to the motion vector detection circuit 26. Motion vector detection circuit 26
Detects a motion amount and direction, that is, a motion vector from the current frame signal from the two-dimensional low-pass filter 24 and the signal (previous frame signal) one frame before from the one-frame delay circuit 25, and supplies it to the motion vector evaluation circuit 27. .

Here, the motion vector detecting method by the motion vector detecting circuit 26 is a method of subdividing an image into blocks of m pixels × n lines (m and n are integers) and detecting the motion vector for each block. Alternatively, a method of detecting a motion vector for each pixel may be used.
A method with excellent detection accuracy may be applied. As an example of the method of detecting the motion vector for each block, 8 pixels × 8
A method is known in which an image is subdivided into blocks of lines and a motion vector is detected by using an iterative gradient method (for example, "Journal of the Television Society, 1991 Vol.
12, pp 1534-1543 ”).

The motion vector evaluation circuit 27 evaluates the accuracy of the detected motion vector and creates a control signal for the noise removal component. That is, a feedback coefficient (multiplication coefficient) in the first-order cyclic filter 200 described later.
And the execution or non-execution of the isolated point removal operation in the isolated point removal signal selection circuit 29. The motion vector evaluation circuit 27 has a detailed configuration shown in FIG. 6 described later.

Further, the luminance signal Y output from the Y / C separation circuit 22 is delayed by the delay circuit 28 for the same time as the processing delay time in the noise removal control circuit 100, and is passed to the isolated point removal signal selection circuit 29. Supplied. The isolated point removal signal selection circuit 29 constitutes an isolated point removal filter circuit together with a 1-field delay circuit 33 and a motion correction circuit 35 described later.

FIG. 5A shows a basic conceptual configuration of the isolated point removal filter circuit, and FIG. 5B shows a specific configuration of the isolated point removal filter circuit of the first embodiment. Is shown.

The isolated point removal filter circuit generally has
As shown in FIG. 5A, the signal 41 of the current field and the signal 4 of the two scanning lines one field before that scanning line are sandwiched.
3 and 45, a delay circuit 42 having a delay time of 1 field-1 line for taking out a total of 3 signals, a delay circuit 44 having a delay time of 1 line, and these 3 signals 41, 43, It is composed of an isolated point removing signal selection circuit 46 for taking out a signal having the second highest level among 45.

The isolated point removal filter circuit of the first embodiment is
As shown in FIG. 5B, the 1-field delay circuit 33
Is applied as a delay circuit 33A (42) having a delay time of 1 field-1 line and a delay circuit 33B (44) having a delay time of 1 line, and for the moving image area. Also, in order to exert the isolated point removal function, the two signals of the signal of the previous field are motion-corrected by the motion correction circuit 35 in accordance with the motion vector and input to the isolated point removal signal selection circuit 29 (46). I am trying.

Unlike the general isolated point removal signal selection circuit 46, the isolated point removal signal selection circuit 29 of the first embodiment is a control signal for instructing whether or not to execute the isolated point removal operation. Are provided from the motion vector evaluation circuit 27 described above. The isolated-point removing signal selection circuit 29 takes out the signal having the second highest level among the three signals and outputs it to the first-order cyclic filter 200 when the control signal instructs the execution of the isolated-point removing operation. When the control signal is output and the non-execution of the isolated point removal operation is instructed, the signal from the delay circuit 28 is passed as it is and output to the first-order cyclic filter 200.

In the motion compensation circuit 35, the motion vector is formed from the signal different from one frame as described above, whereas the signal to be motion compensated is different from the current signal by one field. Therefore, the motion correction is performed in consideration of this point. For example, the motion vector may be halved for motion compensation. If the current field and the previous field are the same frame, the motion is not compensated, and if the current field is a frame different from the previous field, the motion vector of the frame May be used to perform the motion correction.

The first-order cyclic filter 20 of the first embodiment.
0 is to reduce noise based on inter-frame correlation. The first-order cyclic filter 200 includes a multiplication circuit 30,
It is composed of an adder circuit 31, a multiplier circuit 32, two 1-field delay circuits 33 and 34, and a motion correction circuit 36. Therefore, the structure itself uses a motion vector.
This is the same as the conventional noise reduction circuit having the next recursive filter configuration (see FIG. 3).

That is, the current frame signal output from the isolated point removal filter circuit (isolated point removal signal selection circuit 29) is supplied to the noise removal control circuit 1 in the multiplication circuit 30.
00 is applied to the adder circuit 31 by being multiplied by the coefficient from the motion vector evaluation circuit 27, while the motion-corrected previous frame signal output from the motion correction circuit 36 is applied to the noise removal control circuit 100 in the multiplication circuit 32. Of the motion vector evaluation circuit 27 and the addition circuit 31
The noise is reduced by the addition circuit 31 adding these two signals together, and the resultant signal is sent from the output terminal 37 as an output luminance signal. Further, the noise-reduced signal from the adder circuit 31 is delayed by one frame by sequentially passing through the two 1-field delay circuits 33 and 34, and this delayed signal is evaluated by the motion vector of the noise removal control circuit 100. In response to the control signal from the circuit 27, the motion is corrected by the motion correction circuit 36 and is fed back as the previous frame signal as described above.

As described above, the isolated point removal filter circuit and the first-order recursive filter 200 have the same constitutions as the conventional ones, but the control signals for these circuits are different from the conventional ones.

Therefore, the motion vector evaluation circuit 27 for finally producing the control signal will be described in more detail with reference to FIG.

In FIG. 6, the motion vector evaluation circuit 27
Are two subtraction circuits 54 and 56 and a motion correction circuit 55.
A motion amount determination circuit 57, two absolute value conversion / accumulation circuits 58 and 59, a feedback coefficient generation circuit 60,
It is composed of an isolated point removal filter switching circuit 61. As described above, the motion vector evaluation circuit 27 includes the current frame signal 51 from the two-dimensional low-pass filter 24, the previous frame signal 52 from the one-frame delay circuit 25, and the motion vector 53 from the motion vector detection circuit 26. And are given.

The subtraction circuit 54 obtains the difference between the signal 51 of the current frame and the signal 52 of the previous frame, and after the difference is converted to an absolute value by the absolute value conversion / accumulation circuit 58, a block having a predetermined size is obtained. Are accumulated and input to the feedback coefficient generation circuit 60. The output signal E from the absolute value conversion / accumulation circuit 58 is information that reflects the difference between frames when the motion vector is 0 (hereinafter referred to as interframe difference information when the motion vector is 0).

The previous frame signal 52 is processed by the motion correction circuit 55 so that the motion vector 53 can be obtained for each predetermined block.
The difference between the current frame signal 51 and the motion-corrected previous frame signal is obtained by the subtraction circuit 56, and the difference is converted into an absolute value conversion / accumulation circuit 59.
After being converted into an absolute value by the feedback coefficient generation circuit 6
Input to 0. The output signal F from the absolute value conversion / accumulation circuit 59 is information that reflects a difference between frames when motion is corrected (hereinafter, referred to as inter-frame difference information at the time of motion correction).

Interframe difference information E when the motion vector is 0
The predetermined block for obtaining the inter-frame difference information F at the time of motion correction may be arbitrary, that is, it may be a pixel unit or arbitrary n pixels × m lines.
However, since it is related to the evaluation of the motion vector, it is desirable that the block is smaller than the motion vector detection block. For example, if the motion vector detection block is 8 pixels × 8 lines, about 4 pixels × 2 lines. Is preferred.

The motion amount determination circuit 57 is a circuit for determining the size of the detected motion vector, and for example, determines the size of each of the horizontal vector VX and the vertical vector VY. As an example of the determination, when one of the horizontal direction vector VX and the vertical direction vector VY is larger than the threshold value, the motion vector is invalidated, and both the horizontal direction vector VX and the vertical direction vector VY have the same size. The motion vector is validated when is less than or equal to the threshold value. The valid / invalid signal from the motion amount determination circuit 57 is given to the feedback coefficient generation circuit 60 and the isolated point removal filter switching circuit 61. It should be noted that the motion amount determination circuit 57 has a threshold value α for forming the valid / invalid signal given to the feedback coefficient generation circuit 60 and a threshold value for forming the valid / invalid signal given to the isolated point removal filter switching circuit 61. The hold value β may be different.

As described above, the motion amount determining circuit 57 limits the motion vector for the following reason. When the motion vector is large, the frequency band of the image is lowered due to the accumulation effect of the image pickup device, and when the motion is large, the noise becomes visually inconspicuous and the motion in the first-order cyclic filter 200 or the isolated point removal filter circuit is reduced. Even if the corrected previous frame signal is fed back, its effect will be lessened, and conversely, if the motion vector is erroneously detected, the deterioration of the image quality becomes conspicuous, so it was decided to limit the detected motion vector. .

The feedback coefficient generating circuit 60 outputs the output signals E and F of the absolute value converting / cumulating circuits 58 and 59,
The valid / invalid signal from the motion amount determination circuit 57 is used as an input signal to generate a set of multiplication coefficients 62 (62a and 62b) from these signals.

For example, when the valid / invalid signal indicates that the valid / invalid signal is valid, the feedback coefficient generating circuit 60 further includes the inter-frame difference information E when the motion vector is 0 and the inter-frame difference information F when the motion correction is performed. The difference value EF is compared with a certain threshold value γ to see if EF> γ.
In this case, it is determined that the motion vector is correctly determined, and the feedback coefficient 62a is increased (with this, the multiplication coefficient 62b for the current frame signal becomes relatively small),
If E−F ≦ β, the feedback coefficient 62a is reduced (with this, the multiplication coefficient 6 for the current frame signal is reduced).
2b is relatively large).

Since the difference value EF represents the difference in inter-frame correlation between the case where motion correction is performed and the case where motion correction is not performed, a small difference value EF means that the effect of motion correction is small. Therefore, as described above, the feedback coefficient 62a is reduced to reduce the feedback amount of the motion-corrected previous frame signal.

The feedback coefficient generating circuit 60 minimizes the feedback coefficient 62a (for example, 0) when the valid / invalid signal from the motion amount judging circuit 57 indicates invalid, and the first-order cyclic filter 200 is used. In this case, the feedback amount of the previous frame signal in (1) is most limited to prevent the image quality from being deteriorated due to the feedback of a predetermined value or more. At this time, the motion correction circuit 36 may be stopped from giving the motion vector (63) and the motion correction itself may be stopped.

The isolated point removal filter switching circuit 61 is supplied with a valid / invalid signal from the motion amount determination circuit 57 and the above-mentioned difference value EF from the feedback coefficient generation circuit 60. The isolated point removal filter switching circuit 61, when the valid / invalid signal indicates the validity of the motion vector, and the difference value EF is larger than a certain threshold value δ (may be equal to γ). Outputs a control signal 64 for instructing execution of an isolated point removal operation to the isolated point removal signal selection circuit 29 described above,
In other cases, the isolated point removal signal selection circuit 29
Then, the control signal 64 for instructing non-execution of the isolated point removal operation is output. As described above, the isolated point removal signal selection circuit 29 uses the control signal 64 from the motion vector evaluation circuit 27.
In response to the execution of the isolated point removal operation, the signal having the second highest level is selected and output from the three signals, and when the execution of the isolated point removal operation is instructed, the delay circuit 28
The signal from is passed through as it is. The isolated point removal filter switching circuit 61 may stop giving the motion vector (65) to the motion correction circuit 35 and stop the motion correction when the isolated point removal operation is not instructed.

In the above configuration, when a composite video signal is input, the Y / C separation circuit 22 outputs Y / C.
C separation is performed, the luminance signal Y to be subjected to noise reduction is extracted, and this signal is input to the isolated point removal filter circuit via the delay circuit 28, and at the same time, the noise reduction control circuit 10
Input to 0.

In the noise reduction control circuit 100, the motion vector is detected, the motion vector is evaluated, and the isolated point removal filter circuit and the first-order cyclic filter 2 are used.
00 to form a control signal.

In the isolated point removal filter circuit, in accordance with the control signal from the noise reduction control circuit 100, when the isolated point removal operation causes deterioration in image quality, the input signal is passed as it is to the first-order cyclic filter 200. Output,
When the isolated point removal operation works effectively, the isolated point is removed through the selection operation from the three signals and the result is output to the first-order cyclic filter 200.

In the first-order recursive filter 200, when the feedback amount of the motion-compensated previous frame signal is increased in accordance with the control signal from the noise reduction control circuit 100, if the image quality is deteriorated, the current frame signal If the addition ratio is increased and the noise reduction operation is hardly executed and the feedback of the motion-compensated previous frame signal works effectively, the feedback amount of the previous frame signal is increased to reduce the noise and such processing is performed. The finished signal is sent from the output terminal 37 as an output signal of the noise reduction circuit.

Therefore, according to the first embodiment, the motion vector is limited to control the feedback amount of the previous frame signal in the primary recursive filter 200, and the motion vector detection accuracy is evaluated to 1 Since the feedback amount of the previous frame signal in the next recursive filter 200 is controlled, the optimum noise reduction operation can be executed according to the image content such as the still image area, the low speed moving image area, the high speed moving image area, The image quality of the output signal can be improved as compared with the conventional case.

Further, according to the first embodiment, since the motion vector is limited and the motion vector detection accuracy is evaluated to control the execution / non-execution of the isolated point removal operation, the image quality is deteriorated. The isolated point removal operation can be stopped for the moving image area that may cause the noise.

FIG. 7 shows a second embodiment of the noise reduction circuit according to the present invention, and the same or corresponding portions as those in FIG. 1 described above are designated by the same reference numerals.

The second embodiment is a first-order cyclic filter 2
The noise reduction control circuit 100 uses the 1-frame delay circuit (two field delay circuits 33 and 34) in 00, and the 1-frame delay circuit (see the circuit 25 in FIG. 1) is provided in the noise reduction control circuit 100. It is not necessary to provide it.

That is, in the first embodiment, the current frame signal and the previous frame signal required by the noise reduction control circuit 100 are created, but in the second embodiment, they are supplied from the first-order cyclic filter 200 side. It has become so. That is, the output of the adder circuit 31 is supplied as a current frame signal to one input terminal of the motion vector detection circuit 26 and the motion vector evaluation circuit 27 via the first two-dimensional low-pass filter 24A. The previous frame signal, which is the current frame signal delayed by one frame via the two 1-field delay circuits 33 and 34, is the second frame signal.
It is supplied to the other input terminals of the motion vector detection circuit 26 and the motion vector evaluation circuit 27 via the three-dimensional low-pass filter 24B.

According to the second embodiment, the same effect as that of the first embodiment can be obtained, and further, the noise reduction circuit as a whole can be more easily constructed by sharing the one-frame delay circuit. be able to.

FIG. 8 shows a third embodiment of the noise reduction circuit according to the present invention, and the same or corresponding portions as those in FIG. 7 described above are designated by the same reference numerals.

In the third embodiment, the first-order cyclic filter 20 is used.
The motion compensation circuit 55 in the motion vector evaluation circuit 27 is also used as the motion compensation circuit (36) of 0, and the configuration is simplified more than in the second embodiment. Although not shown in FIG. 8, the signal line connected from the motion vector evaluation circuit 27 to the multiplication circuit 32 is a signal line connected to the output terminal of the motion correction circuit 55 of the motion vector evaluation circuit 27 of FIG. Is.

Since the motion compensation circuit 55 in the motion vector evaluation circuit 27 is also used as the motion compensation circuit (36) of the first-order cyclic filter 200, the signal from the motion compensation circuit 55 has a high frequency component. Must also be included. In other words, it is necessary that the signal has the same band as the signal from the isolated point removal signal selection circuit 29. Therefore, in the third embodiment, unlike the second embodiment, the current frame signal and the previous frame signal that do not pass through the two-dimensional low-pass filters 24A and 24B are input to the motion vector evaluation circuit 27. There is.

According to the third embodiment as well, the same effect as that of the second embodiment can be obtained, and further the effect that the noise reduction circuit as a whole can be constructed more easily by sharing the motion correction circuit. be able to.

In each of the above embodiments, the noise reduction circuit for the luminance signal is shown, but the kind of the video signal is not limited to this, and the noise reduction circuit for the color signal (chroma signal, color difference signal or primary color signal). It is also applicable to a noise reduction circuit for a composite video signal. For example, when applied to a noise reduction circuit for a color signal (chroma signal, color difference signal or primary color signal), only a luminance signal is used for motion vector detection, and noise removal is performed for both a luminance signal and a color signal. Is also good.

Further, in each of the above embodiments, the noise reduction control circuit 100 has the two-dimensional low-pass filters 24, 4 and 4.
Although the filters A and 24B are provided, a signal whose frequency is not limited by this filter may be used for detecting a motion vector or the like. Alternatively, the separated chroma signal or the separated color difference signal may be used in the noise reduction control circuit 100.

Further, in each of the above-described embodiments, the control signal generation configuration for the first-order cyclic filter and the control signal generation configuration for the isolated point removal filter circuit are shown in common. The control signal generating structure may be separately provided. For example, the motion vector detection configuration may be separate.

[0062]

As described above, according to the first aspect of the present invention,
In a noise reduction circuit having a first-order cyclic filter configuration using motion vectors, the motion vectors are evaluated, and the ratio of the motion-corrected previous frame video signal fed back to the current frame video signal is changed according to the evaluation result. Since the feedback ratio control means is provided, the image quality of the video signal after the noise reduction processing can be improved more than ever before, regardless of the image content of the video signal.

According to the second aspect of the present invention, in the noise reduction circuit having the isolated point removal filter circuit configuration, the motion vector detecting means for detecting the motion vector of the video signal of the current field video signal one or several fields before. When,
Based on at least the magnitude of the detected motion vector, the isolated point removal signal selecting means sets the isolated point removal signal execution means to the signal selection operation state based on the correlation or the current field video signal passing state. Since the execution control means is provided, the isolated point removal operation can be executed only when the isolated point removal operation is effective image content, and the image quality of the video signal output from the isolated point removal signal selection means is improved as compared with the prior art. be able to.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a noise reduction circuit according to a first embodiment.

FIG. 2 is a block diagram showing a conventional noise reduction circuit having a first-order cyclic filter configuration.

FIG. 3 is a block diagram showing a conventional noise reduction circuit having a first-order cyclic filter configuration using motion vectors.

FIG. 4 is a diagram illustrating the operation of an isolated point removal filter.

FIG. 5 is an explanatory diagram of a configuration of an isolated point removal filter circuit according to the first embodiment.

FIG. 6 is a block diagram showing a detailed configuration of a motion vector evaluation circuit of the first embodiment.

FIG. 7 is a block diagram showing a configuration of a noise reduction circuit according to a second embodiment.

FIG. 8 is a block diagram showing a configuration of a noise reduction circuit according to a third embodiment.

[Explanation of symbols]

25: 1 frame delay circuit, 26: motion vector detection circuit, 27: motion vector evaluation circuit, 29: isolated point removal signal selection circuit, 30, 32: multiplication circuit, 31: addition circuit, 33, 34: 1 field delay Circuit, 35, 36:
Motion compensation circuit, 100: noise reduction control circuit, 200 ...
First-order recursive filter.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Takeo Tsutsui 2-2-1, Jinnan, Shibuya-ku, Tokyo Within Japan Broadcasting Association (72) Yoshiharu Hoshino 2-2-1, Jinnan, Shibuya-ku, Tokyo Within the sending association

Claims (6)

[Claims] 1. A motion compensation is performed on a previous frame video signal using a motion vector, and the motion compensated previous frame video signal is fed back to the current frame video signal at a predetermined feedback ratio to be included in the current frame video signal. A noise reduction circuit having a first-order recursive filter configuration for reducing a noise component having a feedback ratio control means for evaluating the motion vector and changing a feedback ratio according to the evaluation result. Reduction circuit. 2. The feedback ratio control means, the magnitude of the motion vector, and the inter-frame difference information between the motion-corrected previous frame video signal and the current frame video signal,
2. The feedback ratio is changed using the inter-frame difference information between the previous frame video signal and the current frame video signal, which are not motion-corrected, as a parameter.
The noise reduction circuit described in. 3. The feedback ratio control means, when the magnitude of the motion vector exceeds a predetermined value, inter-frame difference information between the motion-corrected previous frame video signal and the current frame video signal, 3. The feedback ratio is limited to a predetermined value irrespective of the value of the inter-frame difference information between the previous frame video signal which is not motion-corrected and the current frame video signal.
The noise reduction circuit described in. 4. A current field video signal and at least one
A target signal extracting means for extracting a plurality of video signals having line correlation in terms of position from the video signals before the field and one of the plurality of extracted signals is selected based on the correlation between these signals. In a noise reduction circuit for removing isolated point noise, which comprises an isolated point removing signal selecting means, a motion vector detecting means for detecting a motion vector for a video signal of the current field video signal one or several fields before, and at least detecting Execution / non-execution of isolated point removal for setting the isolated point removal signal selection means to the signal selection operation state based on the correlation or the passing state of the current field video signal based on the magnitude of the generated motion vector. A noise reduction circuit having a control means. 5. The difference when the isolated point removal execution / non-execution control means does not perform motion correction between the video signals used in the detection of the motion vector and separated by at least one field or more, in addition to the motion vector. The noise reduction circuit according to claim 4, wherein the setting state of the isolated point removal signal selection means is switched using information and difference information after motion correction as parameters. 6. The noise reduction circuit according to claim 4, wherein the past video signal taken out by the target signal taking-out means is motion-corrected.