JPH04354487A - Moving picture coding system and its decoding system - Google Patents

Abstract

PURPOSE:To obtain the moving picture coding system and its decoding system employing the. rate control system with high estimate accuracy of an information quantity. CONSTITUTION:An input digital picture signal is split into plural bands by an analysis filter 130. Coders 140-1-140-n encode a picture signal of each band based on a coding parameter S170. A multiplexer 160 multiplexes an output of each coder and sends the result to a transmission line 180 and gives a coding rate S160 to a rate control circuit 170. The rate control circuit 170 applies linear prediction analysis to the coding rate S160, predicts the information quantity of the picture signal to be coded to decide the coding parameter S170 and to control the said coding rate S160. A moving picture decoder 200 demultiplexes the code from a transmission line 180 for each band to reproduce the picture signal at each band.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture encoding method and its decoding method, and more particularly to a moving picture encoding transmission rate control method for controlling the encoding rate in moving picture encoding.

0002

2. Description of the Related Art Conventionally, this type of technology has been described in, for example, the following documents.

1990 Image Coding Symposium (PC
SJ90), (1990-10) Takishima and Wada, "Rate Control Method in Packet Video Coding", P. 239-
242 As described in the above literature, a video packet communication method using an asynchronous transfer mode network (ATM network) is attracting attention as an exchange transmission method for B-ISDN, which is expected to be a next-generation network. . The packet communication method enables large-capacity communication by collecting encoded audio and images into packets (cells), adding a header indicating the destination and content type, and transmitting and receiving the cells over an ATM network. It is something to do. By taking advantage of the large capacity and asynchronous characteristics, and performing image encoding at a variable rate,
High-quality image communication becomes possible.

However, if the total number of cells exceeds the processing capacity of the ATM network, the overflowing cells will be discarded. Therefore, in the technique of the above-mentioned document, the amount of code (encoding rate) generated from the variable rate video encoder is controlled in consideration of network management, fee setting, and service provision. The configuration will be explained below using FIG. 2.

FIG. 2 is a block diagram of a conventional moving picture encoding device with an encoding rate control function.

This video encoding device has an input terminal 1 for inputting an image signal, and an adder 2 is connected to the input terminal 1.
, discrete cosine transformer (hereinafter referred to as DCT transformer) 3
, and a quantizer 4 are connected. An inverse quantizer 5, a DCT inverse transformer 6, and an adder 7 are connected to the output side of the quantizer 4, and a frame memory 8 and a motion vector detector 9 are connected to the adder 7. . The quantizer 4 also includes a rate control circuit 10 and a variable length encoder 11.
are connected to each other, and an output terminal 12 is connected to them.

Next, the operation will be explained.

When one frame of image signal is input from the input terminal 1, the motion vector detector 9 performs block matching with the image signal in the frame memory 8 and the input signal to obtain a motion vector, and outputs the motion vector. terminal 12
Output from. The adder 2 subtracts a prediction signal S8 considering the motion vector from the input image signal to generate a residual signal S.
Find 2.

The residual signal S2 is transformed by a DCT transformer 3, and the obtained DCT coefficient S3 is quantized by a quantizer 4 based on a step size given from a rate control circuit 10. The quantized coefficient S4 is sent to the variable length encoder 1
1, the data is variable-length coded using a Huffman code, run-length code, etc., and the variable-length code S11 is output from the output terminal 12.

The inverse quantizer 5 inversely quantizes the code based on the step size given from the rate control circuit 10 and outputs the quantized DCT coefficient S5 to the DCT inverse transformer 6. The DCT inverse transformer 6 inversely transforms the quantized DCT coefficient S5 into a residual signal S6. The residual signal S6 is added with the previously subtracted prediction signal S8 in an adder 7 to become a locally decoded signal S7, and the locally decoded signal S7 is stored in the frame memory 8 as a prediction signal S8 of the next frame. be done.

The operation of the rate control circuit 10 will now be explained.

The rate control circuit 10 monitors the amount of code per frame output from the output terminal 12, and calculates the step size Qn of the quantizers 4 and 5 for the frame In to be encoded using the following equation (1). Decide as follows. △Q=a(In-1 -Ifc) Qn =Qn-1 +△Q...(1
) However, △Q; increment value a of Q; sensitivity coefficient In-1; code amount If of previous frame; short-term target information amount = [Io (n-1+N)
-ΣIi]/Ni=1 to n-1 (target amount of information to set the average of n-1+N frames to Io) Ifc; correction value of If=kIo + (1-k)I
f (k<0) Io ; Target amount of encoded information per frame N
;Rate convergence parameter When quantization is performed with the step size Qn of the quantizers 4 and 5 determined by equation (1), the average encoding rate converges to Io. Note that Qn is also output from the output terminal 12.

[0013]

[Problem to be Solved by the Invention] However, in the video encoding transmission rate control method in the video encoding method of the above-mentioned document, the quantization step size Qn is determined by estimating the information amount of the frame In to be encoded. For estimation, a simple linear sum of the code amounts In-1 of past frames is used. Therefore, there is a problem that the estimation of the amount of information is not very accurate, and it has been difficult to solve this problem.

[0014] The present invention provides a video encoding method and its decoding method that solves the problem of the prior art, that is, the accuracy of estimating the amount of information is low in the video encoding transmission rate control method. It is something to do.

[0015]

[Means for Solving the Problems] In order to solve the above problems, a first invention divides an image signal into one or more bands, and applies predetermined encoding parameters to the divided bands. In the video encoding method for encoding, a rate control circuit performs a linear predictive analysis of the encoding rate, predicts the amount of information of the image signal to be encoded, determines the encoding parameter, and performs the encoding. I try to control the rate.

[0016] The second invention is a video encoding method in which an image signal is divided into a plurality of bands and the image of each band is encoded using predetermined encoding parameters. Then, the amount of information of the image signal to be encoded is predicted, encoding parameters are determined for each band, and the encoding rate is controlled.

[0017] In a third invention, the rate control means of the first or second invention is configured as follows. That is, the rate control circuit includes a weighting coefficient calculation means that analyzes the information amount of past frames and calculates a weighting coefficient of the information amount of the past frame based on a predetermined equation; an estimated code amount calculation means for calculating an estimated code amount of a frame to be encoded using the estimated code amount; and an encoding parameter control means for controlling an encoding parameter of the frame to be encoded based on the estimated code amount. .

[0018] The fourth invention is such that, in the video decoding system, an image encoded by the video encoding system of the first or second invention is decoded based on the transmitted code. ing.

[0019]

[Operation] According to the first invention, since the video encoding method is configured as described above, when an image signal is divided into one or more bands and the divided bands are encoded, Then, the rate control circuit performs a linear predictive analysis of the encoding rate,
The amount of information of the image signal to be encoded is predicted, the encoding parameters are determined, and the encoding rate is controlled. Based on the encoding parameters determined by controlling the encoding rate, highly accurate encoding can be performed on the divided bands.

According to the second invention, when dividing an image signal into a plurality of bands and encoding an image in each band, the rate control circuit performs a linear predictive analysis of the encoding rate and determines the image to be encoded. The amount of information in the signal is predicted, coding parameters are determined for each band, and the coding rate is controlled. Based on the encoding parameters determined by controlling the encoding rate, encoding for each band can be performed with high precision.

According to the third invention, the weighting coefficient calculation means calculates the weighting coefficient of the information amount of the past frame. Using this weighting coefficient, the estimated code amount calculation means:
Calculate the estimated code amount of the frame to be encoded. The encoding parameter control means controls encoding parameters based on the estimated code amount. In this way, since the code amount of the frame to be encoded is estimated by linear predictive analysis from the code amount of past frames, the code amount can be estimated with high accuracy.

According to the fourth invention, an image encoded using the video encoding method of the first or second invention is extracted from the transmitted code, for example, by extracting its encoding parameters,
Since encoding is performed based on the encoding parameters, a reproduced image with good image quality can be obtained.

[0023]

Embodiment FIG. 1 is a block diagram of a moving picture encoding/decoding apparatus showing an embodiment of the present invention.

[0024] This device includes a moving image encoding device 100,
The moving picture decoding device 200 is connected to the moving picture decoding device 200 via a transmission line 180.

The video encoding device 100 has a television camera (TV camera) 110 that inputs an input original image, and an analog/digital converter (hereinafter referred to as an A/D converter) on the output side. An analysis filter 130 is connected via 120 . The analysis filter 130 includes A/
This is a circuit that divides the output of the D converter 120 into a plurality of n bands, and a plurality of n encoders 140-1 to 140-1 are connected to the output side of the circuit.
140-n is connected. Each encoder 140-1~
140-n is a circuit that encodes each image signal band-divided by the analysis filter 130 based on the encoding parameter S170, and is configured by DCT intraframe encoding, motion compensated interframe encoding + DCT, etc. .

These encoders 140-1 to 140-n
A multiplexer 160 is connected to the output side of the multiplexer 160, and a rate control circuit 170 and a transmission line 180 are connected to the output side of the multiplexer 160. The multiplexer 160 includes encoders 140-1 to 14
This circuit multiplexes the outputs of 0 to n and sends it to the transmission line 180, and also outputs the encoding rate S160 to the rate control circuit 170. The rate control circuit 170 inputs the encoding rate S160 and controls each encoder 140-1 to 140.
-n encoding parameter S170 and controls the encoding rate S160.

The moving image decoding device 200 uses the transmission path 180
The separator 180 separates the code into the code of each band and extracts the encoding parameter S170, and the output side is provided with a synthesis filter 240 via a plurality of n decoders 220-1 to 220-n. is connected.

Each of the decoders 220-1 to 220-n inputs the code separated into each band by the separator 210 based on the encoding parameter S170, reproduces the image signal of each band, and sends it to the synthesis filter 240. This is a circuit that outputs to DCT.
Intraframe decoding, motion compensated interframe decoding + DCT
It is made up of etc. The synthesis filter 240 is a circuit that synthesizes the outputs of the decoders 220-1 to 220-n and outputs one reproduced image signal, and has a digital/analog converter (hereinafter referred to as D/A) on the output side. (referred to as a converter)2
An image monitor 260 is connected via 50.

Next, the moving image encoding method and its decoding method shown in FIG. 1 will be explained.

In the moving image encoding device 100, the input original image is converted into an electric signal indicating the light density distribution by scanning the image pickup tube of the TV camera 110,
0, the signal is quantized and sampled as a pixel by, for example, 8 bits, converted into a digital image signal, and then sent to the analysis filter 130. The analysis filter 130 converts the digital image signals output from the A/D converter 120 into a plurality of n
The image signal of each band is divided into bands and encoded by an encoder 140-1.
~140-n.

[0031] In each encoder 140-1 to 140-n,
Based on the encoding parameter S170 output from the rate control circuit 170, the image signal of each band is encoded. This encoded signal is multiplexed by multiplexer 160 and sent to transmission path 180. Also, the multiplexer 16
0 outputs the encoding rate S160 to the rate control circuit 170. The rate control circuit 170 controls the encoding rate S1
60 is input, the encoding parameter S170 of each encoder 140-1 to 140-n is determined, and the encoding rate S1
60.

The code sent from the transmission path 180 is separated into codes for each band by a separator 210 in the moving image decoding device 200, and sent to decoders 220-1 to 220-n. At this time, separator 210 extracts encoding parameter S170 from the sent code and provides it to decoders 220-1 to 220-n. Decoder 220
-1 to 220-n reproduce the image signals of each band based on the encoding parameter S170 and send them to the synthesis filter 240.

The synthesis filter 240 synthesizes a plurality of n band-divided image signals into one reproduced image signal. This reproduced image signal is converted into an analog image signal by the D/A converter 250 and displayed on the image monitor 260.

FIG. 3 shows each encoder 140- in FIG.
1 to 140-n. FIG.

This encoder is composed of a motion-compensated interframe predictive encoder, and has an input terminal 141 for inputting an input image signal, which includes a motion vector detector 14.
A frame memory 143 is connected via 2. Further, an adder 144 , a DCT converter 145 , and a quantizer 146 are connected to the input terminal 141 . The quantizer 146 includes an inverse quantizer 147 and a DCT inverse transformer 14.
A frame memory 143, a motion vector detector 142, and an adder 144 are connected to the adder 149. In addition, the quantizer 146
is connected to an output terminal 151 via a variable length encoder 150.

This encoder operates as follows.

First, the input image signal from the input terminal 141 is matched with the image signal of the previous frame stored in the frame memory 143 by the motion vector detector 142 to detect a motion vector. As a result, the predicted signal S143 including the motion vector is subtracted, resulting in a residual signal S144.

The residual signal S144 is processed by a DCT converter 145.
The DCT coefficient S145 is sent to the quantizer 146. The quantizer 146 quantizes the DCT coefficient S145 based on the quantization parameter given from the rate control circuit 170 in FIG. 1, that is, the encoding parameter S170. This quantization coefficient S146 is
The variable length encoder 150 performs variable length encoding using a Huffman code, run length code, etc., and the variable length code S15
0 is output to the output terminal 151.

The quantized coefficient S146 is also sent to an inverse quantizer 147. The dequantizer 147 dequantizes the quantization coefficient S146 based on the encoding parameter S170, and outputs the output transform coefficient S147 to the DCT inverse transformer 148.
send to The output transform coefficient S147 is the DCT inverse transformer 14
8, the DCT inverse transform is performed, and the output transform coefficient S14
8 is the prediction signal S143 previously subtracted by the adder 149
is added and stored in the frame memory 143 as a locally reproduced image signal S149.

The motion vector detected by the motion vector detector 142 is output from the output terminal 151. The quantizer 146 and the inverse quantizer 147 are connected to the rate control circuit 17
Although it is controlled by the encoding parameter S170 starting from 0, this encoding parameter S170 is also output from the output terminal 151.

FIG. 4 shows each decoder 220- in FIG.
1 to 220-n. FIG.

This decoder is constructed of a motion compensated interframe predictive decoder, and is a circuit for decoding the image signal encoded by the encoder of FIG.

This decoder has an input terminal 22 for inputting the code.
1, which includes an adder 2 via a variable length decoder 222, an inverse quantizer 223, and a DCT inverse transformer 224.
25 are connected. A frame memory 226 and an output terminal 227 are connected to the adder 225.

Next, the operation of this decoder will be explained.

The code from the transmission path 180 in FIG.
70 and is variable length decoded by a variable length decoder 222 via an input terminal 221. This variable length decoder 2
The output S222 of No. 22 is dequantized by the dequantizer 223 based on the encoding parameter S170.

Output transform coefficient S223 of the inverse quantizer 223
is inversely DCT transformed by the DCT inverse transformer 224,
The output transformation coefficient S224 is added with a prediction signal S226 in which the motion vector is added in an adder 225, and is stored in the frame memory 226 as a reproduced image signal S225. The obtained reproduced image signal S225 is output from the output terminal 227.

Next, the configuration and operation of the rate control circuit 170 shown in FIG. 1 will be explained with reference to FIG. 5.

FIG. 5 shows the encoding rate S160 and signal-to-noise ratio (SNR) of the variable rate video encoding apparatus shown in FIG.
) is a diagram showing the temporal variation of

As shown in FIG. 5, the encoding rate S160
The SNR is kept constant due to the fluctuation of . The rate control circuit 170 is a circuit that performs linear predictive analysis on this encoding rate curve to estimate the code amount of a frame to be encoded.

The rate control circuit 170 includes weighting coefficient calculation means for analyzing the information amount of past frames and calculating weighting coefficients for the information amount of the past frames based on a predetermined equation; Estimated code amount calculation means for calculating an estimated code amount of a frame to be encoded using
and encoding parameter control means for controlling encoding parameters of a frame to be encoded from the estimated code amount.

For example, it is assumed that the information of the i frame is In, the estimated information amount of the i frame is Iwn, the k-th linear prediction parameter of the i frame is αkn, and the estimation error of the information amount of the i frame is en.

The weighting coefficient calculation means calculates αkn that minimizes en 2 by solving the K-th order simultaneous equations based on the following equation (2). en =In −Iwn=In −Σαk
n・In−k (k=1,…,K)


(2) When the linear prediction parameter αkn is determined, the estimated code amount calculation means calculates the estimated information amount Iwn based on the following equation (3). Iwn=Σαkn・In−k
(k=1,...,K
)

(3) Next, the encoding parameter control means uses the estimated value of equation (3) to calculate, for example, the following equation (
4), controlling the quantization step size QBn,
Control the amount of information. In other words, from the estimated information amount QBn, the encoding parameter S170 of the frame to be encoded is
control. QBn=QBn-1+ΔQB
(Iwn≧I0)


QBn=QB
n-1-ΔQB (
Iwn<I0) ... (4) However, QBn: quantization step size of each band (QBmin<
QB <QBmax) ΔQB: Increment of step size for each band (ΔQB >
0) I0: Target encoding rate In this example, the code amount of the frame to be encoded is estimated by linear predictive analysis from the code amount of past frames, so the code amount estimation accuracy is high and accurate. rate control. Furthermore, in this embodiment, the analysis filter 13
Since subband encoding using 0 is used, encoding distortion can be distributed according to human visual characteristics, deterioration of image quality can be suppressed as much as possible, and the encoding rate can be precisely controlled.

Specifically, ■ΔQB is changed for each band. ■Perform analysis in each band and set target encoding rate for each band. This can be easily realized.

In the above embodiment, an example of subband image encoding using the analysis filter 130 has been described, but various modifications are possible.

For example, in FIG. 1, the analysis filter 130 and multiplexer 160 in the video encoding device 100 are removed, and the separator 21 in the video decoding device 200 is removed.
0 and the synthesis filter 240, it is also possible to implement each of the encoder 140-1 and decoder 220-1 by one encoder 140-1 and one decoder 220-1. As a result, substantially the same effects as those of the above embodiment can be obtained with a simple circuit configuration.

[0056]

[Effects of the Invention] As explained in detail above, according to the first invention, the encoding parameters are determined by performing linear predictive analysis on the encoding rate and predicting the amount of information of the image signal to be encoded. Since the coding rate is controlled, the code amount can be estimated with high accuracy and accurate rate control can be performed.

According to the second invention, the image signal is divided into a plurality of bands, the coding parameters are determined for each band, and the coding rate is controlled. The amount estimation accuracy is good, and appropriate rate control can be performed.

According to the third invention, the weighting coefficient calculating means for calculating the weighting coefficient calculates the weighting coefficient of the information amount of the past frame, and using the weighting coefficient, the estimated code amount calculating means calculates the weighting coefficient of the information amount of the past frame. Calculate the estimated code amount of the frame,
The encoding parameter is controlled by the code amount parameter control means based on the estimated code amount. Therefore,
The code amount of a frame to be encoded can be estimated easily and accurately by linear predictive analysis from the code amount of past frames.

According to the fourth invention, since the image encoded by the video encoding method of the first or second invention is decoded based on the transmitted code, reproduction with good image quality can be achieved. An image is obtained.

[Brief explanation of the drawing]

FIG. 1 is a configuration block diagram of a moving image encoding/decoding device showing an embodiment of the present invention.

FIG. 2 is a configuration block diagram of a conventional video encoding device.

FIG. 3 is a configuration block diagram of the encoder in FIG. 1;

FIG. 4 is a configuration block diagram of the decoder in FIG. 1.

FIG. 5 is a diagram showing temporal fluctuations in the encoding rate and SNR in FIG. 1;

[Explanation of symbols]

100 Video encoding device 130
Analysis filters 140-1 to 140-
n encoder 160 multiplexer 170 rate control circuit 180
Transmission path 200 Video decoding device 210
separator

Claims (4)

[Claims] Claim 1: In a video encoding method in which an image signal is divided into one or more bands and the divided bands are encoded using predetermined encoding parameters, a rate control circuit controls the encoding rate. 1. A moving image encoding method, characterized in that the encoding parameter is determined by linear predictive analysis of the image signal to predict the amount of information of the image signal to be encoded, and the encoding rate is controlled. Claim 2: In a video encoding method in which an image signal is divided into multiple bands and the image in each band is encoded using predetermined encoding parameters, a rate control circuit performs linear predictive analysis of the encoding rate and performs encoding. 1. A moving image encoding method, characterized in that the amount of information of the image signal to be processed is predicted, encoding parameters are determined for each of the bands, and the encoding rate is controlled. 3. The video encoding method according to claim 1, wherein the rate control circuit analyzes the amount of information in past frames, and weights the amount of information in the past frames based on a predetermined equation. a weighting coefficient calculation means for calculating a coefficient; an estimated code amount calculation means for calculating an estimated code amount of a frame to be encoded using the weighting coefficient; and an estimated code amount calculation means for calculating an estimated code amount of a frame to be encoded from the estimated code amount an encoding parameter control means for controlling encoding parameters;
Video encoding method with 4. A moving image decoding method, characterized in that an image encoded by the moving image encoding method according to claim 1 or 2 is decoded based on a transmitted code.