JP3207215B2 - Image coding transmission rate control method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding transmission rate control system.

[0002]

2. Description of the Related Art A moving picture packet communication system using an asynchronous transfer mode network (ATM network) has attracted attention. In the packet communication method, a large-capacity communication is performed by collecting encoded audio and video into packets (cells), adding a header indicating a destination and a type of content, and transmitting and receiving cells on an ATM network. Is what you do. By performing image coding at a variable rate by utilizing the features of large capacity and asynchronous, high quality image communication becomes possible. In this case, if the total number of cells exceeds the processing capacity of the ATM network, overflowing cells will be discarded. Regarding such a problem, the following document discloses a method of controlling the amount of coding information (coding rate) generated from a variable-rate video encoder in consideration of network management, charge setting, and service provision. Is described. Document titles: Takishima, Wada, "Rate control method in packet video coding", 1990 Symposium on Image Coding (PCSJ90), pp. 239-242, 19
October 1990

In this document, the amount of encoded information for one frame output from an encoder output terminal is monitored.
Step size Q of quantizer of frame In to be encoded
n is determined as follows, for example. ΔQ = a (In−1−Ifc) Qn = Qn−1 + ΔQ ΔQ: Q increment a: Sensitivity coefficient In−1: Encoded information amount of the (n−1) th frame If: Short-term target information amount = [Io (n -1 + N) -Σ
Ii] / N, i = 1 to n-1 (target information amount for setting the average of n-1 + N frames to Io) Ifc: Correction value of If = kIo + (1-k) If (k
<0) Io: Target coded information amount for a frame N: Rate convergence parameter When quantization is performed with the quantizer step size Qn obtained by this equation, the average coding rate converges on Io. Was.

[0004]

However, in the above document, when estimating the amount of encoded information of a frame to be encoded and determining a quantization step, a simple linear calculation of the amount of encoded information of a past frame is performed. Since the sum is used, there is a disadvantage that the estimation of the information amount is not very accurate. Therefore, an object of the present invention is to provide a rate control method with high estimation accuracy of the amount of encoded information.

[0005]

According to the present invention, an adaptive filter is provided, and the amount of coded information for one frame is provided as an input to the adaptive filter for each frame period, and the output of the adaptive filter is combined with the output of one frame. Are compared with the target coded information amount, and the quantization step size is changed in a direction to increase or decrease the coded information amount according to the comparison result. Then, the tap coefficients of the adaptive filter are sequentially updated every period of one frame so that the absolute value error between the input and output of the adaptive filter is reduced. Alternatively, the coding information amount of a predetermined band component is provided as an input to the adaptive filter for each period of one frame, and the magnitude relationship between the output of the adaptive filter and the target coding information amount of the predetermined band component is compared. Depending on the result, the quantization step size is changed in a direction to increase or decrease the amount of encoded information. Then, the tap coefficients of the adaptive filter are sequentially updated every period of one frame so that the absolute value error between the input and output of the adaptive filter is reduced.

[0006]

The adaptive filter outputs a value obtained by adding the weight of the coded information of the past frame by tap coefficients.
By successively updating with a system identification algorithm such as the steepest descent method in the direction in which the error between the input and output of the adaptive filter becomes smaller, the adaptive filter estimates the amount of coded information at the time of the frame to be coded from now on. Output. This estimated coded information amount reflects a change in the past information amount of the image itself and a change in the past coding parameter, and is compared with the target coded information amount. Can be expected to converge to the target encoded information amount by changing stepwise for each frame.

[0007]

FIG. 1 is a block diagram showing an embodiment of an encoding apparatus according to the present invention. In FIG. 1, the input original image is T
The scanning of the image pickup tube of the V camera 101 converts the light into an electric signal indicating a light and shade distribution, and the A / D converter 102 performs 8-bit quantization and sampling as a pixel to convert it into a digital image signal. This digital image signal is
The image is divided into four bands by the analysis filter 103 (QMF), and the image signals in each band are encoded by the encoders 104a to 104d, multiplexed by the multiplexer 105, and transmitted to the transmission path 107. . Here, the encoders 104a to 104d are configured by DCT intra-frame encoding, motion compensation inter-frame encoding + DCT, and the like.
The rate control circuit 106 receives the coded information from the multiplexer 105, calculates the amount of coded information per frame in the code amount adder 106a, and
In step 6b, the estimated coded information amount of the current frame is calculated from the past coded information amount, and the step setting circuit 106c calculates the code of the current frame in accordance with the comparison result between the estimated coded information amount and the target coded information amount. Determine the optimization parameters.

[0008] The estimated coded information amount Iwn, which is the output of the adaptive filter 106b, can be expressed by equation (1). Iwn = ΣAkn · In-k (k = 1, 2,..., K) Expression (1) where Akn is the k-th time of the n-th frame of the adaptive filter
The next tap coefficient, In-k, is the encoded information amount of the nk-th frame, and K in this embodiment is 5.
The tap coefficient of the adaptive filter 106b is set for each frame period in such a direction that the error between the input and output of the filter (that is, the estimated error of the coded information amount) En shown in Expression (2) becomes smaller for the next frame. Update. En = In−Iwn = In−ΣAkn · In−k (k = 0, 1,..., K−1) Equation (2) In this embodiment, the steepest descent method using the square of the error En as an evaluation function is employed. Then, it is updated every frame period according to the equation (3). Akn + 1 = Akn + ΔA × sign (En × In−k) (k = 0, 1,..., K−1) Expression (3) where ΔA: increment of tap coefficient, sign: 0 in parentheses
Is 0 when minus; -1 if minus; +1 if plus. As described above, if the update of the tap coefficient is continued for every several frame periods over several tens of frames, the adaptive filter 106b outputs an accurate estimated coded information amount along the change curve of the past coded information amount. Will be.

The step setting circuit 106c determines the quantization step size according to the following equation, and
40a to 40d to control the amount of encoded information. QBn = QBn−1 + ΔQB (Iwn ≧ Io) QBn = QBn−1−ΔQB (Iwn <Io) Equation (3) where QBn: quantization step size of each band (QBm)
in <QB <QBmax), ΔQB: increment of step size set for each band in consideration of human visual characteristics, I
o: Target coding information amount. As described above, the estimated coding information amount Iwn is compared with the target coding information amount Io, and the coding parameter is set so as to increase or decrease the coding information amount according to the result of the comparison. Step (ΔQB) is changed. By performing encoding using the encoding parameters determined in this way, the encoded information amount gradually converges to the target encoded information amount in a stepwise manner.

FIG. 2 is a main block diagram of another encoding apparatus according to the present invention, showing a portion corresponding to one of the four divided bands. In addition, this embodiment employs motion-compensated inter-frame predictive coding, sets a target coded information amount for each band in units of each band, estimates the coded information amount for each band, and A unit change amount of an encoding parameter is set for each band. In FIG.
First, the input image signal from the image signal input terminal 200 is sent to the frame memory 20 by the motion vector detector 201.
The motion vector is detected by matching with the image signal of one frame before the second, and the adder 203 subtracts the prediction signal in consideration of the motion vector to obtain a residual signal.

This residual signal is subjected to DCT conversion by DCT converter 204. The DCT coefficient is calculated by the quantizer 20.
5. Quantization and inverse quantization by the inverse quantizer 206,
The variable length encoder 209 performs variable length encoding. The output transform coefficients of the inverse quantizer 206 are inversely transformed by the DCT by the inverse DCT transformer 207, added with the predicted signal subtracted by the adder 208, and stored in the frame memory 202 as a locally reproduced image signal. The sign of the coefficient and the motion vector are output from the sign output terminal 210. Code amount adder 211, adaptive filter 212, and step setting circuit 2
13 estimates the coding information amount of this band, compares it with the target coding information amount of this band, and determines the quantization parameter of this band according to the magnitude comparison result. This quantization parameter is provided to the quantizer 205 and the inverse quantizer 206, and the amount of encoded information is controlled. The quantization parameter or the unit change thereof is also output to the code output terminal 210.
Output from Note that the quantization parameter determined here is a reference amount in this frame, and is also modified by a preset instantaneous upper limit information amount, but detailed description thereof will be omitted.

[0012]

As described above in detail, according to the present invention, the code amount of a frame to be encoded is reduced by using an adaptive filter whose tap coefficients are updated by a system identification algorithm. , The estimation accuracy of the code amount is good, and appropriate rate control can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram of an encoding device according to the present invention.

FIG. 2 is a block diagram of another encoding apparatus according to the present invention.

[Explanation of symbols]

 103 Analysis filter 104a-d Encoder 105 Multiplexer 106 Rate control circuit 106a Code amount adder 106b Adaptive filter 106c Step setting circuit

────────────────────────────────────────────────── ─── Continuation of the front page (56) References Takishima and Wada, "Rate control method in packet video coding", Image Coding Symposium, 5th Symposium Material (PCSJ90), October 1990, p. 239 −242 (58) Field surveyed (Int.Cl. ⁷ , DB name) H04N ^7/ 24-7/68 JICST file (JOIS)

Claims (2)

(57) [Claims] 1. An encoding information amount for one frame is provided as an input of an adaptive filter for each period of one frame, and a magnitude relationship between an output of the adaptive filter and a target encoding information amount for one frame is compared. According to the comparison result, in the direction of increasing or decreasing the amount of encoded information,
Tap coefficients of the adaptive filter are successively updated every period of one frame so that a quantization step size is changed and an absolute value error between input and output of the adaptive filter is reduced. Image coding transmission rate control method. 2. An encoding information amount of a predetermined band component is provided as an input of an adaptive filter for each frame period, and a magnitude relationship between an output of the adaptive filter and a target encoding information amount of the predetermined band component is determined. Then, in accordance with the comparison result, the quantization step size is changed in a direction to increase or decrease the amount of encoded information, and an absolute value error between input and output of the adaptive filter is reduced. A tap coefficient of the adaptive filter is sequentially updated every cycle of one frame.