JPS61173592A - Interframe encoding system - Google Patents

Abstract

PURPOSE:To reduce deterioration in picture quality and to obtain a high- compressibility system by separating a motion picture signal into a motion area and a still area when an image signal is transmitted digitally, and quantizing roughly predicted errors of picture elements in the still area and also quantizing one image plane finely at every constant period. CONSTITUTION:An adder 5 adds the quantized predicted error signal to a prediction signal from a frame memory 6 and supplies the result as a local decoded signal to the frame memory 6. A quantization switch 701 outputs a quantization switching signal which normally selects rough quantization for a part determined as the still area with a motion/still decision signal supplied from a motion/still separating circuit 2 through a signal line 201 and also outputs a quantization switching signal which selects fine quantization for the whole of the still area when a timing signal is sent from a timer 702. Thus, the rough quantization is normally performed by utilizing the motion/still decision signal and buffer occupancy and the whole still area is quantized finely upon occasion.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to digital transmission of image signals, and particularly relates to high-efficiency bandwidth transmission using an interframe coding method that performs different quantization for each motion and static region of an input moving image signal. Regarding compressed transmission technology.

(Prior Art and Problems) Conventionally, as an interframe coding method that separates motion and static of an input video signal and applies different quantization characteristics to each motion and static region, there is a method described in Japanese Patent Application No. 169011/1989 [Moving Image Signal As described in ``Predictive Coding Apparatus'', a predictive coding method is known that applies a quantization characteristic that has a large dead zone (a range where the output is zero relative to the input) to a stationary region. However, with this method, the size of the dead zone changes suddenly between the moving area and the static area, which leaves behind a trail of dirt on the screen after the object moves, and this area in particular becomes the static area. Since the dead zone is enlarged, there is a drawback that dirt remains forever. Furthermore, since the selection of quantization characteristics is usually made in units of multiple lines, subtle differences in gradation may be noticeable in the static area after a scene change at the boundary where the characteristics have changed.

(Objective of the Invention) An object of the present invention is to provide an encoding method that causes less deterioration in the image quality of a decoded image, especially in a still region, and that can obtain a high compression rate.

(Structure of the Invention) The present invention separates a moving image signal into a moving area and a still area in advance, and normally performs coarse quantization on prediction errors for pixels included in the still area, but finely quantizes at least one screen. This is an interframe encoding method characterized by performing quantization.

(Principle of the Invention) In the present invention, first, a moving image signal is separated into a moving region and a still region. Several methods of separation are already known. For example, as described in Japanese Patent Application No. 59-194110 [Moving Picture Signal Dynamic Separator], the screen is divided into blocks of a certain size, and the absolute value of the frame difference value of each pixel in the block is calculated within the block. There is a method in which the motion and static of the block is determined by adding the sum and comparing the result of the addition with a predetermined threshold value. Alternatively, in a method called the gradient method, a pixel-by-pixel motion vector is determined from the luminance gradient within a frame and a frame difference value, and a set of pixels for which this motion vector is not zero can be defined as a motion region. In the present invention, any method of separating motion and static may be used. Next, using FIGS. 1 and 2, a method of performing dense quantization in an arbitrary frame only for pixels included in a still area will be explained. FIG. 1 shows the results of motion and static separation performed using any of the methods described above. 1st
It is assumed that the shaded area in each frame in the figure is determined to be a stationary area by motion and static separation. Generally, coarse quantization is performed on the diagonally shaded static region of each frame, thereby suppressing information generated by noise in the static region.

Next, dense quantization is performed on the entire diagonally shaded stationary region of an arbitrary frame, for example, the n-th frame. A method for performing dense quantization will be explained using FIG. Normally, coarse quantization is performed on the portion determined to be a stationary area by motion/static judgment.For example, fine quantization is applied to the stationary area in the n-th frame, and coarse quantization is applied to the stationary area between the next frame and the P frame. . Then, dense quantization is performed on the static region of the n-th frame. In this manner, dense quantization is performed on the stationary region at appropriate intervals. If the quantization of the static area remains coarse, the dynamic area and the static area will change due to image movement or scene changes.

By performing dense quantization on the entire static area, information generated by noise in the static area can be suppressed, and the information generated by the noise in the static area can be suppressed, and the screen smudges that occur at the boundaries between the moving and static areas can sometimes remain. It is possible to remove screen smudges at boundaries and static areas after a scene change. Furthermore, once the quantization is done finely, even if coarse quantization is performed afterwards, screen smudges will not occur in still areas unless there is movement or a scene change. In this way, an encoding system with a high compression rate can be obtained.

(Example) An example of the present invention will be described in detail using FIG. Third
The figure shows one embodiment of an encoder using the present invention. A digitized moving image signal input through a signal line 101 is supplied to a delay circuit 1 and a motion/static separation circuit 2. The motion/static determination signal calculated by the motion/static separation circuit 2 is supplied to the encoding control circuit 7 via the signal line 201. The moving image signal whose delay has been adjusted by the delay circuit 1 enters the subtracter 3. Subtraction '63 calculates a frame difference based on the predicted signal delayed by one frame time from the frame memory, and sends this to the quantization circuit 4 as a predicted error signal. When quantizing the prediction error signal, the quantization circuit 4 switches between coarse and fine quantization for the static region, for example, as shown in FIG. 2, using a quantization switching signal from the encoding control circuit 7. Take control.

For the dynamic region, quantization determined based on, for example, the occupancy state of the buffer memory 11, which will be described later, is performed as in the conventional case. Next, the prediction error signal quantized by the quantization circuit 4 is supplied to an adder 5 and a code converter 8.

Adder 5 adds the quantized prediction error signal and the prediction signal from frame memory 6, and supplies the result to frame memory 6 as a locally decoded signal. The code converter 8 variable-length encodes and compresses the prediction error signal supplied from the quantization circuit 4.

When performing variable length encoding, the code converter 8 performs variable length encoding according to which quantization characteristic is selected by the quantization switching signal from the encoding control circuit 7, and converts the variable length code into a variable length code. is output to the multiplexing circuit.

The code converter 9 encodes as a mode code which quantization characteristic is selected based on the quantization switching signal from the encoding control circuit 7, and outputs the code to the multiplexing circuit. Multiplexing circuit 1
0 multiplexes the mode code and variable length code and outputs it to the buffer memory 11.

The buffer memory 11 matches the encoding speed and the transmission path speed so as to keep the transmission speed of the transmission path 1101 constant. Further, buffer occupancy indicating the occupied state of the buffer memory 11 is supplied to the encoding control circuit 7 via 1102.

Next, the encoding control circuit 7 will be explained with reference to FIG. The encoding control circuit 7 is comprised of a quantization switch 701 and a timer 702, as shown in FIG. A timer 702 supplies a timing signal to the quantization switch at an arbitrary period, for example, in units of several frames. Further, the timer 702 is connected to the signal line 1102 from the buffer memory 11.
The amount of generated information is monitored by the buffer occupancy supplied via the buffer, and the timing signal is not output when there is a large amount of generated information. The quantization switch 701 outputs a quantization switching signal that normally selects coarse quantization to the portion determined to be a stationary region by the motion/static determination signal supplied from the motion/static separation circuit 2 via the signal line 201, and the timer 70
Only when a timing signal is sent from 2, a quantization switching signal is output for selecting dense quantization for the entire stationary region. In this way, depending on the motion/static determination signal and the buffer occupancy, coarse quantization is normally performed, and sometimes fine quantization is performed over the entire stationary region. At this time, if there are many occurrences (that is, when the signal line 1102 indicates that the buffer occupancy is large), coarse quantization is performed instead of fine quantization.

Also, instead of supplying the signal 1102 indicating buffer occupancy to the timer 702, the signal 1102 is supplied to the quantization switch 701, and when the buffer occupancy becomes large, the timer 702 supplies the signal 1102. Almost the same effect can be obtained even if the configuration is configured so that the switching signal is ignored.

Next, the decoder will be explained with reference to FIG. The buffer memory 12 matches the compression-encoded signal sent from the encoder at a constant speed through the transmission path 1201 with the decoding speed, and outputs the signal to the code inverse exchanger 13.14. The code inverse exchanger 13 extracts only the mode code from the multiplexed compressed code and decodes it. The decoded signal (representing the quantization characteristic) is sent to the code inverse exchanger 1 by 1301.
4 and is used to control quantization switching. The code inverse exchanger 14 decodes the variable length code excluding the mode code among the multiplexed compressed codes based on this quantization characteristic, and outputs the decoded prediction error signal to the adder 15. . The adder 15 adds the decoded prediction error signal from the code inverse exchanger 14 and the prediction signal from the frame memory 16, and supplies the decoded signal to the frame memory 16. Ru.

(Effects of the Invention) As explained in detail above, coarse quantization is performed on the static region to suppress information generated by noise in the static region, while quantizing the static region gradually from medium to fine on a frame-by-frame basis. By performing quantization, it is possible to remove screen smudges in static areas, which have been a problem in the past, by adding a small amount of information, and an interframe coding method with a high compression rate can be realized. When the present invention is put to practical use in this way, its effects are extremely large.

[Brief explanation of drawings]

FIGS. 1 and 2 are diagrams for explaining the present invention in detail. In the figure, 1...Delay circuit 2...Dynamic/static separation circuit 3...Subtractor 4...Quantization circuit 5...Adder 6... Frame memory 7... Encoding control circuit 8...
... Code converter 9 ... Code converter 10.
...Multiplexing circuit 11...Buffer memory 12...Buffer memory 13...Sign inverse converter Engineering 4...Sign inverse converter 15...- Addition i 16...
・Frame memory. Two Patent Attorneys Uchihara's House□ Kabuto□ House□ ・ ・ ・ ・

Claims (1)

[Claims] When performing interframe predictive coding on a moving image signal such as a television signal, the moving image signal is separated into a moving area and a still area, and prediction errors for pixels included in the still area are coarsely quantized. 1 every fixed period
An interframe encoding method that is characterized by performing dense quantization on the screen.