JPH09224250A - Level fluctuation detector and image quality improving device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect the level fluctuation in a decoded frame signal. SOLUTION: A decoded frame signal is converted into a pseudo optical signal by a gamma characteristic circuit 101 and a mean luminance signal of the frame is calculated by a 1-frame mean calculation circuit 102. The obtained mean luminance signal is delayed sequentially by one frame at a 1-frame delay circuit 103, and the 1-frame mean calculation circuit 102 outputs the mean luminance signal of a frame of a B picture type in a timing when the 1-frame delay circuit 103 outputs the mean luminance signal of the frame of the P picture type. Then a difference between the mean luminance signal of the frame of the B picture type from the 1-frame mean calculation circuit 102 and the mean luminance signal of the frame of the B picture type is outputted from a difference circuit 107 to output the mean luminance difference between the frame of the B picture type and the frame of the P picture type.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a level fluctuation detecting apparatus for detecting a level fluctuation of a decoded image of a signal coded by a coding method using bidirectional prediction, and a level fluctuation detecting apparatus for detecting the level fluctuation. The present invention relates to an image quality improving device that improves decoded image quality based on level fluctuation.

[0002]

2. Description of the Related Art As a method for improving the quality of a displayed image, several kinds of proposals have been made so far.

A first method is to remove noise components on an electric signal supplied to a display device by adaptively using a spatiotemporal filter. For example, "a direct current addition type noise reducer is used. Proposal and Development of Noise Reducer for MUSE Decoder Video Processing "(Izumi et al., Journal of TV Society, vol.48, N
O.12, pp1553-1564 (1994)) is known.

The second method is for flicker, and when a fluorescent lamp is used to illuminate the subject, a TV is used.
(television) This is to detect flicker in an electrical signal output from a camera by detecting a temporal variation of a signal that gives a spatially same position, correct the temporal variation, and remove a flicker.
Such a method is described in, for example, "Fluorescent lamp flicker removal device for television signals" (Otsuka et al., NHK STRL, vol.
27, No. 12, pp.501-506) are known.

The third method is to correct variations in gamma characteristics between different types of monitors. This is because, for example, the gamma characteristic is considerably different between the projection type and the direct-view type and affects the gradation characteristic on display, so that the difference in the gamma characteristic is corrected by a process called "inverse gamma: re-gamma". As such a method, "Examination of gradation characteristics and image quality of high-definition display" (Kondo, Kanazawa, 1992)
The one described in the annual conference of the Television Society of Japan 5-8) is known.

Although flicker occurs due to various causes, if it can be detected on an electric signal, it can be detected and corrected by the second method, but if it is not,
Could not be corrected. For example, when the spatial resolution difference between images of temporally consecutive frames is reflected in the average luminance level difference, the correction cannot be performed. Normally, in a display device such as a CRT, since there is a gamma characteristic, there is no linear relationship between the levels of an electric signal and an optical signal. Therefore, even if the electric signal level has the same average value, a high frequency signal exists. This can result in different average values at the optical signal level.

An example of such a temporal change in spatial resolution that causes flicker is the display of a decoded image of a coding method using bidirectional prediction. As data compression methods, there are known a data compression method that is completed within the image regardless of the preceding and succeeding images on the time axis and a compression method by unidirectional prediction that performs prediction in time series. The former compression encoding structure is an I picture (Intra coded Pictu).
res) is a P picture (P
redictive coded Pictures). The I picture is an intra-frame coded image obtained by directly performing DCT (discrete cosine transform) of the video signal, has the lowest compression rate, and can be reproduced without referring to other pictures when decoding. The P picture is a picture which is forward predicted only from a temporally past picture (I picture or P picture) and inter-frame coded. As shown in FIG. 4A, I pictures are periodically arranged to support special reproduction functions such as random access and fast-forward reproduction. The GOP (Group
of pictures).

However, since this one-way prediction can only predict from a frame preceding in time, it is impossible to predict the future, for example, the uncovered portion that appears as the object moves. It was

As a data compression method capable of predicting the future, MPEG (Moving Picture Experts Group)
Video encoding methods are known. An image coded by this coding method is a B picture (Bidirectionally Predictive code) in addition to an I picture and a P picture.
d Pictures). B-pictures are those that are predicted from both the past and the future and are inter-frame coded. MPEG
In the method, a coded image is configured by repeating an image unit represented by M in FIG. 4B, which includes one P picture or I picture and one or more predetermined number of B pictures.

In the coding example shown in FIG. 4B, there are two B pictures between the I picture and the P picture. In a B picture in which bidirectional prediction is used, a predicted image from a temporally previous frame, a predicted image from a temporally backward frame, and an average image (interpolation prediction) of these two predicted images are selected. Used for. Therefore, more accurate prediction is performed as compared with a P picture that can only perform prediction from the previous frame. The uncovered part described above can be predicted from a picture that is behind in time, and the problem of the uncovered part does not substantially exist in bidirectional prediction.
In addition, when interpolation prediction is used, the compensation accuracy is higher than the accuracy of motion compensation in one direction, and even when there is uncompensable motion such as deformation, a predicted image with a high degree of phase approximation can be obtained. (See Figure 5).

Therefore, in many pictures, the probability that interpolation prediction is used for B pictures is high. However, in the interpolative prediction, the average of the images with the shifted phases is taken, so that the spatial resolution is lower than that of the predicted image from one direction. When motion compensation is ideally performed for motion that does not involve deformation, the phase difference between the two images for averaging is at most the magnitude of compensation accuracy. However, in reality, the accuracy of compensation may be larger than the original due to imperfections in the motion estimation, and in the case of an image accompanied by deformation, a large phase difference may exist. Therefore, the spatial resolution of the predicted image of the B picture is lower than the spatial resolution of the P picture,
The degree of the decrease is not uniform because it moves as described above and depends on the deformation.

In bidirectional prediction, B pictures are not used for prediction of other frames, unlike P pictures, so that distortion of B pictures does not affect other frames. In addition, since the prediction efficiency of bidirectional prediction is high as described above, the amount of information required to achieve the same amount of distortion for B pictures is smaller than that for P pictures.
Therefore, the allocation information amount is reduced for B pictures,
Efficient information amount control can be performed by increasing the allocation to I pictures and P pictures accordingly.

[0013]

However, at the same time, due to such information amount control, the information amount is not allocated to the transmission of the high frequency component of the B picture, and as a result, the resolution of the B picture becomes lower than that of other picture types. It will be.

Generally, since a display device such as a CRT has a gamma characteristic, there is no linear relationship between the electric signal level and the optical signal level, and the average value differs at the optical signal level. As a result, even if the average value is the same at the electrical signal level, if there is a high-frequency signal, the spatial resolution difference between the images of the temporally consecutive frames is changed to the average luminance level difference,
The average luminance level of the optical signal fluctuates with time, causing flicker.

An object of the present invention is to solve the above problems and to provide a level fluctuation detecting apparatus capable of detecting a flicker level.

Another object of the present invention is to solve the above problems and to provide an image quality improving apparatus capable of suppressing flicker and improving decoded image quality.

[0017]

[Means for Solving the Problems]

1) The level fluctuation detecting apparatus according to the present invention predicts one unidirectional interframe or interfield predictive coded image in which only temporally past signals are used as prediction signals, and temporally past and future signals. The unit is repeated with one or more predetermined number of bidirectional frame-to-frame or inter-field predictive-coded images as signals, and periodically,
First signal processing means for performing non-linear processing having electro-optical conversion characteristics on a decoded image signal in which the unidirectional inter-frame or inter-field predictive encoded image is replaced with an intra-frame or intra-field encoded image, The decoded image signal obtained by the processing of the first signal processing means is encoded into an intra-frame or intra-field encoded image, a unidirectional interframe or interfield predictive encoded image, and a bidirectional interframe or interfield predictive encoded. First average value processing means for performing an average value processing in a predetermined range in each frame or field of the image, and a predetermined number of bidirectional frames included in the image unit obtained by the first average value processing means By the arithmetic mean value of the average value processing output of each inter-field or inter-field predictive coded image and the processing of the first average value processing means The difference between the unidirectional inter-frame or inter-field predictive coded image is output, and the arithmetic mean value and the intra-frame or intra-field coded image obtained by the process of the first average value processing means are output. And a difference output means for outputting the difference between

2) In the level fluctuation detecting device described in 1) above, the first signal processing means performs gamma correction conversion on each decoded image signal.

3) In the image quality improving apparatus according to the present invention, the difference output output by the level fluctuation detecting apparatus described in 2) above is averaged by the first average value processing means described in 1) above. Bidirectional bidirectional interframe or inter-field predictive coded images are added to the predetermined number of decoded image signals, and the same number of first addition units as the predetermined number, and inverse gamma for each of the signals obtained by the first addition unit. The same number of second signal processing units as the predetermined number for performing the correction conversion, and the average value processing of the decoded image signal of the bidirectional inter-frame or inter-field predictive coded image within a predetermined range within a frame or field are performed. Second average value processing means to perform, and average value processing of the predetermined number of bidirectional interframe or interfield predictive coded images obtained by the processing of the second average value processing means. The same number of second difference output means as the predetermined number for outputting the difference between the output and the signal obtained by the signal processing by the second signal processing means, and the difference output by the second difference output means, The present invention is characterized by comprising a predetermined number of second adding means for adding the predetermined number of decoded image signals of the bidirectional inter-frame or inter-field predictive coded image for each pixel.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION

<First Embodiment> FIG. 1 shows a first embodiment of the present invention. The first embodiment is a case where the predetermined number of bidirectional inter-frame or inter-field predictive coded images included in an image unit is one. That is, to the image quality improving apparatus shown in FIG. 1, a decoder (not shown) supplies a decoded frame having a picture type of P picture and B picture at a cycle of 2 frames and a frame of an I picture type at a predetermined cycle. There is. This I picture type frame is supplied in place of the P picture type frame at the time of refreshing. The decoder refers to the picture type in the internal processing and can retrieve the picture type signal from the decoder. The output timing signal is a control signal that turns ON at the output timing of a P picture type frame or an I picture type frame.

The image quality improving apparatus in the case where the P picture and B picture type frames are supplied in this way is composed of a level fluctuation detecting section and a level fluctuation suppressing section.

The level fluctuation detecting section includes a gamma characteristic circuit 101 as a first signal processing means, a 1-frame average calculating circuit 102 as a first average value processing means, a 1-frame delay circuit 103 as a difference output means, and Difference circuit 1
07 and a latch 108.

The gamma characteristic circuit 101 converts a decoded frame signal from a decoder (not shown) into a pseudo optical signal. The 1-frame average calculation circuit 102 is set to 1 for 1 frame based on the pseudo optical signal from the gamma characteristic circuit 101.
The average luminance of the frame is calculated and an average luminance signal is generated. 1 frame delay circuit 103
The average luminance signal from the frame average calculation circuit 102 is delayed by one frame. Difference circuit 10
Reference numeral 7 outputs the difference between the average luminance signal from the 1-frame average calculation circuit 102 and the average luminance signal from the 1-frame delay circuit 103, and the average luminance difference between the B picture type frame and the P picture type frame, that is, flicker. The average luminance difference signal (electrical signal) representing the level is output. The latch 108 responds to the output timing signal by
The average luminance difference signal from the difference circuit 107 is latched once every two frames and supplied to the outside.

The level fluctuation suppressing section includes an adding circuit 109 as a first adding means, an inverse gamma characteristic circuit 111 as a second signal processing means, a 1-frame average calculating circuit 113 as a second average value processing means, It has a difference circuit 122 as a second difference output means, a 1-frame delay circuit 116, a switch 118, and an adder circuit 120 which constitute a second addition means.

The addition circuit 109 adds the average luminance difference signal from the difference circuit 107 and the average luminance signal of the B picture type frame from the one-frame average calculation circuit 102. The inverse gamma characteristic circuit 111 is an addition circuit 109.
The signal from is subjected to inverse gamma conversion to calculate the corrected average electric signal level. The 1-frame average calculation circuit 113 calculates the average luminance of the frame for each frame from the decoded frame signal from the decoder and generates a frame average electric signal. Difference circuit 12
Reference numeral 2 denotes an average electrical signal of the B picture type frame from the 1-frame average calculation circuit 113 and the inverse gamma characteristic circuit 1
The difference from the corrected average electric signal from No. 11 is taken to generate an average value correction difference electric signal. Switch 11
8 is closed once every two frames by the output timing signal. The one-frame delay circuit 116 delays the decoded frame signal by one frame. Addition circuit 1
Reference numeral 20 denotes a B picture type decoded frame signal from the 1-frame delay circuit 116 and an average value correction differential electrical signal obtained from the difference circuit 122 via the switch 118 for each pixel to obtain a B picture type An image quality improving decoded frame signal is generated.

Next, the operation of the level fluctuation detecting section will be described. A decoded frame signal from a decoder (not shown) is converted into a pseudo optical signal by the gamma characteristic circuit 101, and the average luminance signal of the frame is converted into one frame by the one frame average calculation circuit 102 based on the obtained pseudo optical signal. Calculated once. The obtained average luminance signal is sequentially delayed by one frame by the one-frame delay circuit 103. At the timing when the 1-frame delay circuit 103 outputs the average luminance signal of the P picture type frame, the 1-frame average calculation circuit 102 outputs the average luminance signal of the B picture type frame.

Next, the average luminance signal of the B picture type frame from the 1-frame average calculation circuit 102 and 1
The difference from the average luminance signal of the P picture type frame from the frame delay circuit 103 is output by the difference circuit 107, and the average luminance difference between the B picture type frame and the P picture type frame, that is, the flicker level is output. It This flicker level is latched by the output timing signal output once every two frames.
Latched.

Next, the operation of the level fluctuation suppressing section will be described. The decoded frame signal from a decoder (not shown) is also input to the 1-frame average calculation circuit 113 and the 1-frame delay circuit 116 of the level fluctuation suppressing unit. Based on the decoded frame signal input to the 1-frame average calculation circuit 113,
1 for 1 frame by 1 frame average calculation circuit 113
Degrees, the frame average electrical signal level is calculated.

Further, the average luminance difference signal and the average luminance signal are input from the above-mentioned level variation detecting section to the level variation suppressing section. That is, the average luminance difference signal is the difference circuit 107.
From the 1-frame average calculation circuit 102 to the addition circuit 109. Then, the average luminance difference signal and the average luminance signal are added by the adding circuit 109 to generate a corrected average luminance level, and the corrected gamma characteristic circuit 111 calculates the corrected average electrical signal level based on the corrected average luminance level from the adding circuit 109. Calculated. Then, the difference circuit 122 generates a difference electric signal for mean value correction representing a difference between the corrected average electric signal level from the inverse gamma characteristic circuit 111 and the frame average electric signal level of the B picture type from the one-frame average calculation circuit 113. Is output.

On the other hand, by the 1-frame delay circuit 116,
After the B picture type frame is delayed by one frame and supplied to the adding circuit 120, the switch 118 is closed for one frame time. Therefore, at the timing when the flicker signal is output from the level fluctuation detection unit,
The differential electrical signal for average value correction from the differential circuit 122 is supplied to the addition circuit 120 via the switch 118, and the differential electrical signal for average value correction is of the B picture type delayed by one frame by the one frame delay circuit 116. The decoded frame signal and the addition circuit 120 add each pixel,
The corrected B picture type image quality improvement decoded frame signal is output.

In the P picture type frame, since the switch 118 is closed only once every two frames,
It is output from the 1-frame delay circuit 116 without correction.

Therefore, it is possible to perform the DC shift corresponding to the average level variation on the decoded image signal having the reduced spatial resolution, and to suppress the flicker due to the variation of the spatial resolution.

<Second Embodiment> FIG. 2 shows a second embodiment of the present invention.
An embodiment will be described. The second embodiment is a case where the predetermined number of bidirectional inter-frame or inter-field predictive coded images included in an image unit is two. This is an example applied to a decoded image of a typical MPEG signal. FIG. 2 shows an image quality improving device and an MPEG decoder, and the image quality improving device is composed of a level fluctuation detecting section and a level fluctuation suppressing section.

As shown in FIG. 3, the MPEG decoder 200 supplies decoded frames in the order of P picture, B picture, and B picture at a 3-frame cycle. At the time of refreshing, an I picture type frame is supplied instead of the P picture type frame. The MPEG decoder 200 refers to the picture type in the internal processing and can take out the picture type signal.

The level fluctuation detecting section includes a gamma characteristic circuit 201 as a first signal processing means, a 1-frame average calculating circuit 202 as a first average value processing means, a 1-frame delay circuit 203 as a difference output means, and 204,
Adder circuit 205, 1/2 circuit 206, and difference circuit 20
7 and a latch 208.

The gamma characteristic circuit 201 converts the decoded frame signal from the MPEG decoder 200 into a pseudo optical signal. The 1-frame average calculating circuit 202 calculates the average luminance of the frame once based on the pseudo light signal from the gamma characteristic circuit 201.
The 1-frame delay circuit 203 is a 1-frame average calculation circuit 2
The average luminance signal from 02 is delayed by one frame. The adder circuit 205 uses the average luminance signal from the 1-frame average calculation circuit 202 and the 1-frame delay circuit 20.
The average luminance signal from 3 is added. The 1/2 circuit 206 is for obtaining an arithmetic mean from the average luminance signals obtained by adding by the adding circuit 205. The 1-frame delay circuit 204 delays the average luminance signal from the 1-frame delay circuit 203 by 1 frame. The difference circuit 207 outputs an average brightness difference signal representing the difference between the arithmetic average value of the average brightness signals from the 1/2 circuit 206 and the average brightness from the 1-frame delay circuit 204.
The latch 208 latches the average luminance difference signal from the difference circuit 207 once every three frames according to the signal from the MPEG decoder 200.

The level fluctuation suppressing section includes one adder circuit 209 as the first adding means, one inverse gamma characteristic circuit 211 as the second signal processing means, and one frame average calculation as the second average value processing means. The circuit 213, the one-frame delay circuit 214, the one difference circuit 222 as the second difference output means, the other addition circuit 210 as the first addition means, and the other inverse gamma characteristic as the second signal processing means. The circuit 212, the other difference circuit 223 as the second difference output means, and the switch 217 forming the second addition means.
And 218 and 1-frame delay circuits 215, 216,
221 and addition circuits 219 and 220. Each of the first adding means, the second signal processing means, the second difference outputting means, and the second adding means is provided two by two, corresponding to the number of B-picture frames included in the three-frame cycle. The 1-frame delay circuit 221 is shown in FIG. 2 for convenience of explanation, and is not essential as a constituent element of the second adding means.

The adder circuit 209 adds the average brightness difference signal from the difference circuit 207 and the average brightness signal from the 1-frame delay circuit 203. Inverse gamma characteristic circuit 2
Reference numeral 11 is for calculating the corrected average electric signal level based on the signal from the adder circuit 209. The adder circuit 210 adds the average brightness difference signal from the difference circuit 207 and the average brightness signal from the one-frame delay circuit 202. The inverse gamma characteristic circuit 212 calculates a corrected average electric signal level based on the signal from the adder circuit 210.

The one-frame average calculation circuit 213 is an MPEG
Based on the decoded frame signal from the decoder 200, the average luminance of the frame is calculated once for each frame. The one-frame delay circuit 214 delays the frame average electric signal from the one-frame average calculation circuit 213 by one frame.

The difference circuit 222 is the one-frame delay circuit 21.
The difference between the average electric signal of the B picture type frame from 4 and the corrected average electric signal from the inverse gamma characteristic circuit 211 is output to generate a difference electric signal for average value correction. The difference circuit 223 takes the difference between the average electric signal of the B picture type frame from the one-frame average calculation circuit 213 and the corrected average electric signal from the inverse gamma characteristic circuit 212 to generate an average value correction difference electric signal. It is a thing.

The one-frame delay circuit 215 delays the decoded frame signal from the MPEG decoder 200 by one frame. The adder circuit 219 adds the B-picture type decoded frame signal from the 1-frame delay circuit 215 and the average value correction differential electric signal obtained from the difference circuit 223 via the switch 217 for each pixel. The one-frame delay circuit 216 delays the signal from the adder circuit 219 by one frame. 220
Is an adder circuit, and B from the 1-frame delay circuit 216
Picture type decoded frame signal and difference circuit 222
From the average value correcting differential electric signal obtained via the switch 218 from the above. The 1-frame delay circuit 221 delays the signal from the adder circuit 220 by 1 frame.

Next, the operation of the level fluctuation detecting section will be described. The decoded frame signal from the MPEG decoder 200 is converted into a pseudo optical signal by the gamma characteristic circuit 201, and based on the obtained pseudo optical signal, the average luminance of the frame is calculated by the 1-frame average calculation circuit 202 to be 1 per frame. Is calculated. The obtained average luminance signal is sequentially delayed by one frame by the one-frame delay circuits 203 and 204. At the timing when the 1-frame delay circuit 204 outputs the average luminance signal of the P picture type frame, the 1-frame average calculation circuit 203 outputs P
The average luminance signal of the B picture type frame (hereinafter referred to as the first B picture type frame) following the picture type frame is output, and the 1 frame delay circuit 202 outputs the average luminance signal of the first B picture type frame. The next B picture type frame (hereinafter, the second B
An average luminance signal of a picture type frame) is output.

Next, the average luminance signal of the second B picture type frame from the 1 frame average calculation circuit 202 and the average luminance signal of the first B picture type frame from the 1 frame delay circuit 203 Adder circuit 20
5 and the 1/2 circuit 206 adds two B's.
The luminance average of the picture type frames is calculated. 1
The difference between the brightness arithmetic average value from the / 2 circuit 206 and the average brightness of the P picture type frame from the 1-frame delay circuit 204 is output by the difference circuit 207, and the B picture type frame and the P picture type frame are output. The average luminance difference signal, that is, the average luminance difference signal representing the flicker level is output.

The timing at which such calculation is established is
This is the time when the average luminance signal of the P picture type frame is output from the 1-frame delay circuit 204, and occurs once every 3 frames. This timing signal shown in FIG. 3 is supplied from the MPEG decoder 200 to the latch 208. The flicker signal from the difference circuit 207 is the latch 2
It is supplied to the outside via 08.

Next, the operation of the level fluctuation suppressing section will be described. The decoded frame signal from the MPEG decoder 200 is
1-frame average calculation circuits 213 and 1 of the level fluctuation suppression unit
It is input to the frame delay circuit 215.

Based on the decoded frame signal input to the 1-frame average calculation circuit 213, the 1-frame average calculation circuit 213 calculates the average luminance of the frame once per frame, and outputs the frame-average electric signal. .
The obtained frame average electric signal is 1 frame delay circuit 2
14 delays one frame.

Further, the level fluctuation suppressing section outputs the average brightness difference signal from the difference circuit 207 and the average brightness of the first B picture type frame from the one frame delay circuit 203 to the level fluctuation detecting section. The signal and the average luminance signal of the second B picture type frame from the 1-frame average calculation circuit 202 are input. Then, the average luminance difference signal from the difference circuit 207 and the average luminance signal of the first B picture type frame from the 1-frame delay circuit 203 are added by the adding circuit 209, and the average luminance signal of the first B picture type is added. A corrected average luminance signal is generated that represents the corrected average luminance level of the frame. Also, the difference circuit 2
The average brightness difference signal from 07 and the average brightness signal of the second B picture type frame from the 1-frame average calculation circuit 202 are added by the adder circuit 210 to correct the second B picture type frame. A corrected average luminance signal is generated that represents the average luminance level.

Based on the corrected average luminance signal from the addition circuit 209, the inverse gamma characteristic circuit 211 calculates the corrected average electric signal level of the first B picture type frame. The difference circuit 222 outputs the difference (average value correction difference) between the corrected average electrical signal level from the inverse gamma characteristic circuit 211 and the frame average electrical signal level of the first B picture type from the one-frame delay circuit 214. To be done. On the other hand, the inverse gamma characteristic circuit 212 calculates the corrected average electric signal level of the second B picture type frame based on the corrected average luminance signal from the addition circuit 210. The difference (average value correction difference) between the corrected average electric signal level from the inverse gamma characteristic circuit 212 and the frame average electric signal level of the second B picture type from the one-frame average calculation circuit 213 is calculated by the difference circuit 223. Is output.

On the other hand, the first B picture type frame is delayed by one frame by the one-frame delay circuit 216 and started to be supplied to the addition circuit 220. Further, in synchronization with this supply start, the one-frame delay circuit 215 is used. Two
The switch 218 and the switch 217 are simultaneously closed for the time of one frame after the first B picture type frame is delayed by one frame and supplied to the adding circuit 219. Therefore, at the timing when the average luminance difference signal representing the flicker level is output from the level fluctuation detection unit,
The differential electric signal for average value correction from the differential circuit 222 is supplied to the adding circuit 220 via the switch 218, and is added to the first B picture type decoded frame signal delayed by one frame by the one frame delay circuit 216. Is
B picture type frames are corrected. Since the switches 217 and 218 are closed only once in three frames, the first B picture type decoded frame signal is input to the 1-frame delay circuit 216 without being added by the adding circuit 219. . On the other hand, at the timing when the average luminance difference signal representing the flicker level is output from the level fluctuation detection unit, the average value correcting differential electric signal from the differential circuit 223 is supplied to the adding circuit 219 via the switch 217 and delayed by one frame. The circuit 215 adds the second B picture type decoded frame signal delayed by one frame and corrects the B picture type frame. At this timing, the 1-frame delay circuit 221 outputs a P picture type frame.
This P picture type frame is a switch 217,
Since the 218 is closed only once in 3 frames,
It is output from the 1-frame delay circuit 221 without correction. Therefore, flicker due to variations in spatial resolution is reduced.

Even when the predetermined number of bidirectional inter-frame or inter-field predictive-coded images included in an image unit is three or more, the image quality improving apparatus according to the present invention can be used as in the second embodiment. It goes without saying that it can be configured.

[0051]

As described above, according to the present invention,
With the configuration as described above, it is possible to remove or reduce flicker, which is subjectively noticeable and is image quality deterioration, and thus it is possible to improve the image quality of a decoded image to be displayed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a relationship between an output timing signal and a picture type.

FIG. 4 is an explanatory diagram for describing unidirectional prediction and bidirectional prediction as a temporal prediction structure.

FIG. 5 is an explanatory diagram for explaining interpolation prediction in bidirectional prediction.

[Explanation of symbols]

 101 gamma characteristic circuit 102,113 1-frame average calculation circuit 103,116 1-frame delay circuit 107,122 differential circuit 108 latch 109,120 adder circuit 111 inverse gamma characteristic circuit 118 switch

Claims (3)

[Claims] 1. One unidirectional inter-frame or inter-field predictive coded image in which only temporally past signals are used as prediction signals, and one or more one-way type inter-field or inter-field predictive encoded images in which temporally past and future signals are used as predicted signals. The unit is repeated with a predetermined number of bidirectional inter-frame or inter-field predictive coded images as a unit, and the one-directional inter-frame or inter-field predictive coded images are periodically coded within the intra-frame or intra-field code. A first signal processing means for performing non-linear processing having an electro-optical conversion characteristic on the decoded image signal replaced with the digitized image, and the decoded image signal obtained by the processing of the first signal processing means in a frame or in a field. A coded image, a unidirectional interframe or interfield predictive coded image,
First average value processing means for performing an average value processing in a predetermined range within each frame or field of a bidirectional frame-to-frame or inter-field predictive coded image, and the image unit obtained by the first average value processing means A predetermined number of bidirectional frames included in the inter-frame or inter-field predictive coded image, the arithmetic mean value of the respective average value processing outputs, and the unidirectional frame obtained by the processing of the first average value processing means or A difference output for outputting the difference between the inter-field predictive coded image and the difference between the arithmetic mean value and the intra-frame or intra-field coded image obtained by the processing of the first average value processing means. And a level fluctuation detecting device. 2. The level fluctuation detecting apparatus according to claim 1, wherein the first signal processing means performs gamma correction conversion on each decoded image signal. 3. A bidirectional inter-frame or inter-field predictive coded image in which the difference output output by the level fluctuation detecting apparatus according to claim 2 is averaged by the first average value processing means according to claim 1. Number of first addition means to be added to each of the predetermined number of decoded image signals, and a predetermined number to perform the inverse gamma correction conversion on the signals obtained by the first addition means. Two signal processing means; a second average value processing means for performing an average value processing on a decoded image signal of the bidirectional interframe or interfield predictive coded image in a predetermined range within a frame or field; 2 Average value processing output of the predetermined number of bidirectional inter-frame or inter-field predictive coded images obtained by the processing of the average value processing means, and the second signal processing means The same number of second difference output means as the predetermined number for outputting the difference from the signal obtained by the signal processing, and the difference output by the second difference output means for the predetermined number of bidirectional frames or fields. An image quality improving apparatus comprising: a predetermined number of second adding means for adding each pixel to a decoded image signal of an inter prediction coded image.