JPS63309082A - Picture encoding transmission equipment - Google Patents

Abstract

PURPOSE:To improve the quality of reproduced picture by arranging a space filter which smooths a reproduced picture signal and adaptively controlling the smoothing characteristic of the space filter in frame units or block units on the basis of a threshold, motion vector, and block discrimination information. CONSTITUTION:A threshold 17 which is determined for each frame in accordance with the quantity of information generated at the time of preceding frame encoding, vector information 25 which indicates the extent of motion of a block as the encoding object, and block discrimination information 10 which indicates whether a block is encoded as the effective block or not are inputted to a space filter as parameters of adaptive control of the filter. The space filter 26 strongly smooths the signal if the threshold is high, and the filter 26 weakly smooths it if the threshold is low. Smoothing is performed to eliminate the quantization noise if block decision in formation 10 is effective, and smoothing is stopped to avoid high band attenuation of a still area if it is ineffective. Thus, the quality of reproduced picture is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an improvement of an image encoding and transmitting apparatus, particularly an image encoding and transmitting apparatus used for television communications such as video conferences and video telephones.

[Prior Art] Fig. 4 is a block tM diagram of a conventional image encoding and transmitting device. In the figure, (1) represents an analog image signal, and (2)
is an A/D converter, (3) is a digital image signal, (4)
is a block divider that divides one frame of digital image signal into multiple blocks, (5) is an input block signal divided into blocks of a certain size, and (6) is a block divider that performs motion compensation on the previous frame reproduced image. predicted image signal, (7)
is a differentiator, (8) is an inter-frame difference signal, (9) is a comparator, (10) is block determination information, (11) is a block encoding circuit, (12) is block encoding information, (13)
is the transmission buffer memory, (14) is the encoded transmission information, (
15) is the buffer storage information amount, (16) is the encoding control circuit, (17) is the threshold value, (18) is the block decoding circuit,
(19) is the block-decoded reproduced inter-frame difference signal, (20) (adder), (21) is the reproduced image signal, (
22) is a frame memory, (23) is a previous frame reproduced image signal, (24) is a motion compensation circuit, and (25) is motion vector information.

Next, the operation will be explained. When encoding the first frame at the start of transmission, the predicted image signal is forced to O, and the input block signal (5) passes through the subtractor (7) as it is, is block encoded and decoded, and the reproduced image is stored in the frame memory ( 22). At this time, by forcibly setting the threshold value (10) to O, all blocks within one frame are encoded.

In encoding the second and subsequent frames, the difference between the human block signal (5) and the predicted image signal is calculated, and block encoding is performed on this interframe difference signal (8). The predicted image signal (6) is obtained by performing motion compensation on the input block signal (5) using the previous frame reproduced image signal (23).

An example of motion compensation is shown below. Input block signal (5)
The block size is MxN (M, N are positive integers), the position of the block in the frame is ■, and the person block signal is 5-(S,).

...S,l (K-MXN), the position of the reference block in the previous frame reproduced image signal is V'', and the reference block signal is 5-ts''..., S7,1.

Also, the reference vector is expressed as v-(v, v) in Table-X
y, V −−V+v, −X ≦V ≦Xo1−
--Ox-Y≦vS Yo.

y However, X and Y are positive integers of a certain size.

Here, using the pattern matching distortion D-Σ Is -3'' 1 1 , t 1 1 , the reference vector with the smallest pattern matching distortion among the reference vectors (2Xo×2Yo) is determined as the motion vector ■ Then, the reference block signal at the position V−−V+1n−−vmin is set as the block signal at the position V in the predicted image signal.The motion compensation circuit (24) performs the above operation for all blocks in one frame. In addition to outputting a predicted image signal (6), motion vector information (25) representing a motion vector is also output.

Motion compensation has the effect of suppressing interframe difference signal components.

In the comparator (9), mXn (m, n are positive integers, M is a multiple of m,
It is determined whether the block is valid or invalid in units of blocks of size (N is set to be a multiple of n), and block determination information (10) representing the determination result is output. An example of a block determination method is shown below.

Difference signal of 1 block (block size mXn) (7)
When ε-(ε,...S11 (J!-m x n)) is compared with the threshold (17) Th, the block is set to d≧Th
When d<Th, the block is judged to be a valid block, and when d<Th, it is judged to be an invalid block.

The block encoding circuit (11) performs block encoding based on the block determination information (10) when the block is a valid block and outputs block encoding information (12), and outputs block encoding information (12) when the block is an invalid block. No block encoding information (12) is output. Similarly, based on the determination signal, the block decoding circuit decodes the block encoding information (12) when the block is a valid block and outputs the reproduced inter-frame difference signal (19), and when the block is an invalid block, it Output a vector. This reproduced inter-frame difference signal (19) and the preliminarily i1! The +1 image signal (6) is added by an adder (20) to become a reproduced image (21) of the current frame, which is written into the frame memory (22).

On the other hand, the motion vector information (25), the block determination signal (10), and the block encoding information (12) are temporarily stored in a transmission buffer memory (13) and sent out at a constant transmission rate. At this time, in order to suppress the delay of the reproduced image on the receiving side and prevent the transmission buffer memory from overflowing, the amount of information stored in the buffer (13) is smoothed. Buffer accumulation information! 11 When (15) is large, the encoding control circuit (16) increases the threshold value (17) Th to reduce the number of effective blocks in one frame and the encoded information m written to the transmission buffer memory (13). , conversely, when the amount of buffer accumulated information is small, the threshold value (17)
By lowering the value, the amount of information written to the buffer memory (13) is reduced to Jj9.

[Problems to be Solved by the Invention] Since the conventional image encoding and transmitting device is constructed as described above, the amount of transmitted information Ω increases when there is a large change in the image between moving area frames or during intense image transmission. Therefore, the threshold becomes high, and even blocks where the difference signal distortion value is large in the comparator are treated as invalid blocks, resulting in the problem of sticking noise caused by residual components with large amplitudes in the invalid blocks. .

Since there is a large amount of differential signal information in the encoded value of the effective block that has become an effective block in the dynamic domain value, the quantization characteristics become coarse.
There was a problem that coding noise occurred.

Due to the sticking noise and coding noise, the quality of the reproduced image in the frame memory deteriorates, resulting in a problem in that the coding efficiency of subsequent frames decreases.

This invention has been made to solve the above-mentioned problems, and it is an object of the present invention to obtain an image encoding and transmitting device that can suppress encoding noise and sticking noise in a moving region, and can locally control the reproduced image quality. With the goal.

[Means for Solving the Problems] The image encoding and transmitting device according to the present invention arranges a spatial filter immediately before a frame memory, and applies the smoothing characteristics of the filter to a threshold value, motion vector information, and block determination information. The system performs adaptive control based on the threshold value and transmits the threshold value.

[Operations] The image coding and transmitting device according to the present invention suppresses coding noise in the moving region and noise γ1 [sound] by controlling the spatial filter smoothing characteristics in the interframe coding loop on a frame-by-frame and block-by-block basis. 1, and the reproduced image quality improves. Furthermore, good reproduced image quality with high spatial resolution can be obtained in the still area.

[Embodiment] Hereinafter, a preferred embodiment of the present invention will be described with reference to the drawings. In FIG. 1, (26) is a spatial filter that smoothes the reproduced signal, and (27) is a spatial filter output signal. In addition, in FIG. 1, the same parts as those in the conventional example in FIG.

Next, the operation will be explained. The reproduced image signal (21) obtained by the same conventional means is input to a spatial filter (26) and subjected to smoothing processing, and the spatial filter output signal, that is, the smoothed reproduced image signal (27) is converted into a frame. written to memory (22).

FIG. 2 shows the arrangement of input signal samples for the two-dimensional spatial filter on the screen when a two-dimensional spatial filter is used as the spatial filter. smoothed,
Pixel sample value of interest to be output
Assuming G and H, the smoothing process is performed according to the following equation.

10-(A 10C1F + H) < a t + a 2 + a s -1) However, X is the smoothed target pixel sample value, that is, the filter output signal sample value, and a +821a3 is the value for controlling the smoothing characteristic. It is a coefficient.

The larger the coefficient a1 is, the weaker the smoothing effect is, and the smaller the coefficient a1 is, the stronger the smoothing effect is.

An example of an adaptive control method for spatial filter smoothing characteristics will be described below.

As a parameter for adaptive control of the spatial filter (26),
A threshold value (17) determined for each frame according to the amount of information generated during encoding of the previous frame, motion vector information (25) indicating the amount of motion of the block to be encoded, and block determination information indicating whether or not it has been encoded as a valid block. (10)
is input to the filter.

When the threshold value is high, the spatial filter (26) performs strong smoothing because inappropriate pattern sticking is likely to occur, and when the threshold value (17) is low, smoothing is performed weakly to speed up image convergence.

Regarding motion vectors, when the moving distance is large, smoothing is strengthened to suppress high-frequency components of the inter-frame difference signal, and when the moving distance is small, smoothing is weakened.

Furthermore, when the block determination information (10) is valid, smoothing is performed to remove quantization noise, and when it is invalid, smoothing is stopped to avoid high-frequency attenuation in the stationary region.

The above control is not carried out independently of each parameter, but determines the filter characteristics by relating them to each other.

An example of control is shown in FIG.

In the control example shown in Figure 3, smoothing is performed even on invalid blocks when the movement distance of the motion vector is large to prevent inappropriate patterns from remaining, and it is effective when the motion vector is stationary. Adaptive control is performed to stop smoothing even for blocks and to hasten the convergence of stationary parts.

Therefore, when decoding on the receiving side, the threshold value (17) is used to control the filter characteristics, so the threshold value (17) is transmitted.

Although FIG. 1 shows an example in which the spatial filter (26) is provided immediately before the frame memory (22), the same effect can be obtained even if the spatial filter (26) is provided immediately after the frame memory (22).

[Effects of the Invention] As described above, according to the present invention, the image encoding and transmitting device includes a spatial filter for smoothing a reproduced image signal,
Since the smoothing characteristics of the spatial filter are adaptively controlled on a frame-by-frame and block-by-block basis based on the threshold, motion vector information, and block determination information, encoding noise and sticking noise in the moving region are suppressed, and Good reconstructed image quality with high spatial resolution can be obtained in the region, and image quality can be adaptively controlled.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an image encoding and transmitting device according to an embodiment of the present invention, and FIG. 2 is an explanatory diagram showing the pixel arrangement of a spatial filter of the image encoding and transmitting device according to an embodiment of the present invention. FIG. 3 is an example of spatial filter characteristic control, and FIG. 4 is a block diagram showing the configuration of a conventional image encoding and transmitting apparatus. In the figure, (9) is a comparator, (10) is block determination information, (17) is a threshold value, (21) is a reproduced image signal, (2
4) is a motion compensation circuit, (25) motion vector information, (2)
6) is a spatial filter, and (27) is a smoothed reproduced image signal. Note that the same reference numerals in the figures indicate the same or equivalent parts.

Claims (1)

[Scope of Claims] A block divider that divides a plurality of pixels of an image input signal at close positions on an image into blocks, and creates and outputs a vector signal for each block; a motion compensation unit that detects a motion vector indicating a motion of the input vector signal and generates a predicted image signal that is motion-compensated from a reproduced signal of a previous frame stored in a frame memory based on the motion vector; a subtracter that generates a difference signal with the image signal; and calculates an evaluation value indicating the magnitude of the difference signal, compares the evaluation value with a threshold, and if the calculated value is within the range of the threshold. It is determined that the current input block and the block of the predicted signal are the same, the block determination information is set as invalid information, only the block determination information (invalid information) and the motion vector information are encoded and sent to the transmission buffer, and the block determination information (invalid information) and the motion vector information are encoded and sent to the transmission buffer. If the evaluation value is outside the range of the threshold value, the block determination information is set as valid information as a block to be transmitted if the evaluation value is outside the range of the threshold value, and the comparison unit sends the encoded information of the difference signal and the block determination information (valid information) to the transmission buffer; and the motion vector. a block decoding unit that decodes the information, the block determination information, and the encoded information of the effective block to generate a decoded signal; an adder that adds the decoded signal and the prediction signal to generate a reproduced signal; The image encoding and transmitting apparatus includes: a frame memory that stores the predicted signal for each frame; and a threshold control unit that controls the threshold based on the amount of information of the previous frame temporarily stored in the transmission buffer memory. a spatial filter that smoothes the reproduced signal; the threshold value;
It is characterized by comprising means for adaptively controlling the smoothing characteristic value of the spatial filter on a frame-by-frame and block-by-block basis based on the magnitude of the motion vector and the block determination information, and means for transmitting the threshold value. Image encoding and transmission equipment.