JPH02206293A - Method and device for processing of moving image signal - Google Patents

Abstract

PURPOSE:To suppress wasteful information, to decrease a generated information quantity and to make smooth the motion of a subject to be an attention point by fixing a subject part and a background part in a screen beforehand, and making the band width of the background part narrower than the band width of the subject part. CONSTITUTION:A control circuit 1 fixes the pattern of the background beforehand, and accumulates it to a ROM 13. The background pattern is synchronized with an inputted moving image signal, applied to a time filter, and the band width of a time filter 2 is switched. For the background part, by using a characteristic that the amplitude degree of an output to an input is much smaller than one, strict band restriction in time is executed. For the subject part, by using a characteristic that the amplitude degree of the output to the input is at one or below and approximate to one, relatively loose band restriction is executed. Thus since the wastefull information generated due to the motion of the background at the point other than the attention point can be suppressed, as a result, the motion of the subject to be the attention point can be made smooth.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method and apparatus for processing a moving image signal, and particularly to a method and apparatus for processing a moving image signal that can suppress useless information.

[Conventional technology]

In conventional video signal encoding, a noise reducer is applied to the input video signal in order to prevent the encoding efficiency from decreasing due to noise contained in the input video signal.
Encoding was performed after removing noise.

[Problem to be solved by the invention]

In conventional encoding of video image signals, even when the subject of interest is not moving, a large amount of information is generated due to noise or large movements in the background, and encoding is often stopped. , the movement of the subject in focus became discontinuous, which was very unsightly.

In such cases, noise reducers were powerless.

An object of the present invention is to provide a moving image signal processing method and apparatus that can significantly suppress information generated from backgrounds other than points of interest, suppress the generation of useless information, and solve the above-mentioned problems. There is a particular thing.

[Means to solve the problem]

The present invention is a method of processing a moving image signal using a temporal filter, and is characterized by predetermining a subject part and a background part within a screen, and making the bandwidth of the background part narrower than the bandwidth of the subject part. do.

Further, the moving image signal processing device of the present invention stores a plurality of predetermined background patterns, selects one of the background patterns according to a first instruction, and displays a screen using the human-powered moving image signal as a reference. means for outputting a second instruction indicating a subject portion and a background portion in the image; and a means for obtaining the information.

[Operation] In conventional low focus rate video encoding, when the amount of generated information becomes large, encoding is temporarily stopped to match the encoding speed and the transmission path speed. However, with this method, discontinuity occurs in the encoded image when encoding is stopped, resulting in unnatural motion, which is very unpleasant to the eyes.

Therefore, in the present invention, by applying a time filter whose bandwidth is narrower than that of the focused subject part to a predetermined background part, unnecessary noise caused by movement of the background other than the focused subject is eliminated. By suppressing information to reduce the amount of generated information and removing discontinuities in encoded images due to stopping of encoding, the movement of the subject, which is the point of interest, is smoothed.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of a moving image signal processing device of the present invention. This processing device stores a plurality of predetermined background patterns, selects one of the background patterns according to a first instruction, and indicates a subject part and a background part in the screen based on the input video signal. A control circuit 1 constitutes a means for outputting a second instruction, and a means for obtaining a moving image signal having a plurality of bandwidths and whose bandwidth is varied with respect to an input moving image signal according to the second instruction. It has a temporal filter 2 and an encoder 3 that encodes a moving image signal with a changed bandwidth.

FIG. 2 shows an example of the configuration of such a control circuit 1, which is composed of a synchronous separator 11, a counter 12, and a ROM (read only memory) 13.

This control circuit 1 predetermines a background pattern as shown by diagonal lines in FIG. 4 and stores it in a ROM (read only memory) 13. In videophones and the like to which low-vintage video encoding is applied, people are the main focus of attention, so a background pattern as shown by diagonal lines in FIG. 4(A) is prepared. Alternatively, if only the facial part is to be noticed, a pattern such as that shown in FIG. 4(B) may be used. In this way, a plurality of background patterns are prepared and selected depending on the situation. The bandwidth of the temporal filter 2 is switched by applying this background pattern to the temporal filter in synchronization with the input moving image signal.

FIG. 3 shows an example of the configuration of the time filter 2,
It is composed of a subtracter 21, a ROM 22, an adder 23, and a frame memory 24.

The input moving image signal is sent to the frame memory 2 by a subtracter 21.
Subtraction is performed with the signal from one frame ago given from 4. The frame difference signal obtained by the subtracter 21 is R
given to OM22. Characteristics as shown in FIG. 5 are written in the ROM 22 in advance, and the characteristics of the time filter are switched in accordance with a filter switching signal given from the control circuit 1.

For the background portion, by using a characteristic in which the amplification degree of the output with respect to the input is much lower than 1 as shown in FIG. 5(A), strong band limitation is performed in terms of time.

For the object part, relatively light band limitation is performed by using a characteristic in which the amplification degree of the output with respect to the input is less than or equal to 1 and close to 1, as shown in FIG. 5(B). ROM22
In addition to this characteristic, it is also possible to use a nonlinear characteristic in which the amplification degree decreases as the frame difference value increases.

The frame difference signal to which the output of the ROM 22 has been suppressed is added to the signal of one frame before by an adder 23, and becomes a moving image signal to which band limitation has been added. This band-limited moving image signal becomes the output signal of the temporal filter and is also applied to the frame memory 24, where it is used for filtering the next frame.

Next, an embodiment of the moving picture signal processing method of the present invention will be described based on the moving picture signal processing apparatus shown in FIGS. 1 to 5.

The input moving image signal is supplied to the control circuit 1 and the temporal filter 2 via line 100 in FIG.

An input moving image signal supplied via line 100 is applied to sync separator 11 . The sync separator 11 separates sync from the human-powered moving image signal and generates a frame reset signal and a clock for resetting at the beginning of a frame. The frame reset signal generated by the sync separator 11 is supplied to the counter 12 via line 14.

The clock generated by the synchronous separator 11 is also connected to the line 15.
is supplied to the counter 12 via.

The counter 12 is reset at the beginning of the frame by the frame reset signal supplied from the sync separator 11, counts up by the clock supplied from the sync separator 11, and outputs a value indicating the coordinate within the screen. The coordinate value of the output of the counter 12 is supplied to the ROM 13.

As mentioned above, a plurality of background patterns, such as those shown by diagonal lines in FIG. 4, are written in advance in the ROM 12, and are switched depending on the situation. The area indicated by diagonal lines in Figure 4 is defined as the background, and the address of the background coordinates is “0°°
, and write °“1°” in the address corresponding to the subject area.

When focusing on the entire subject, the fourth
A signal for selecting the pattern shown in FIG. 4(A) is supplied to the ROM 13, and when attention is mainly focused on the facial part, a signal for selecting the pattern shown in FIG. 4(B) is supplied to the ROM 13. R.O.
M13 selects the background pattern according to the background pattern selection signal supplied via line 110, and counter 1
Read out the previously written values according to the coordinate values supplied from 2.

The output signal of the ROM 13 is supplied to the temporal filter 2 via a line 120 as a selection signal for the temporal filter 2.

In FIG. 3, the input moving image signal is supplied to the subtractor 21 of the temporal filter 2 via line 100. The subtracter 21 performs subtraction between the human-powered moving image signal and the signal supplied from the frame memory 24 one frame before, and obtains a frame difference signal. The frame difference signal obtained by the subtracter 21 is supplied to the ROM 22.

As mentioned above, input/output related values as shown in FIG. 5 are written in the ROM 22 in advance. And ROM22
When the selection signal of the temporal filter 2 supplied from the control circuit l via the line 120 is "0" indicating the background, the characteristic shown in FIG. Add restrictions and suppress information in the background. When the selection signal of the temporal filter 2 indicates an object at "1" degree, by selecting the characteristics shown in Fig. 5(B) and adding a relatively light band limit, it is possible to prevent sudden movements caused by the movement of the object part. It also suppresses the generation of information. The filtered frame difference signal output from the ROM 22 is sent to the adder 2.
3.

The adder 23 receives the filtered frame difference signal supplied from the ROM 22 and the frame memory 24.
A moving image signal that has been subjected to background processing is obtained by adding the signal from the previous frame supplied from . The background-processed moving image signal obtained by the adder 23 is supplied to the frame memory 24 and is also supplied to the encoder 3 via a line 230 as an output signal of the temporal filter 2. The frame memory 24 delays the background-processed moving image signal supplied from the adder 23 by one frame time, and outputs it to the subtracter 21 and adder as a signal for background processing of the next frame. 23.

The encoder 3 encodes the background-processed video signal supplied from the temporal filter 2 by using correlation between screens or correlation within a screen, and encodes the video signal while matching the speed of the transmission path. Sends the encoded signal to the receiving side.

As described above, according to this embodiment, the background part is determined in advance,
By significantly limiting the bandwidth of the background, it is possible to significantly reduce unauthorized information due to movement in the background, and to code still regions of the background clearly. Furthermore, by applying a relatively light band limit to the object part that is the point of interest, it is possible to naturally encode the movement while suppressing the rapid generation of information in the object part.

〔Effect of the invention〕

As explained in detail above, by using the video signal processing of the present invention, it is possible to suppress useless information generated due to movement of the background other than the point of interest, and as a result, the movement of the subject, which is the point of interest, can be suppressed. Can be made smooth. Therefore, when the present invention is put into practical use, the effect of improving image quality in video encoding is extremely large.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an embodiment of the moving image signal processing device of the present invention, FIG. 2 is a diagram showing FIG. 1, FIG. 3 is a diagram showing the following, FIG. It is a diagram. 1... 2... 3... 11... 12... 13, 22... 21... 23... 24. ... A diagram showing an example of a background pattern, an example of the configuration of the time filter shown in Figure 1, which shows an example of the characteristics of the time filter, and a control circuit showing an example of the configuration of the control circuit shown in the figure.Time filter encoder synchronous separator counter. OM subtracter adder frame memory

Claims (2)

[Claims] (1) A method of processing a moving image signal using a temporal filter, which is characterized by predetermining a subject part and a background part within the screen, and making the bandwidth of the background part narrower than the bandwidth of the subject part. How to process moving image signals. (2) A plurality of predetermined background patterns are stored, one of the background patterns is selected according to a first instruction, and a second pattern indicating the subject part and background part in the screen is selected based on the input video signal. the second instruction; and means for obtaining a moving image signal having a plurality of bandwidths and changing the bandwidth with respect to the input moving image signal according to the second instruction. A processing device for moving image signals.