JPH0575987A - Method for detecting action area of dynamic image signal and high efficiency coder - Google Patents

Abstract

PURPOSE:To obtain the coder of an efficient dynamic image. CONSTITUTION:In a high efficiency coder to compress and code the television signal of a dynamic image, the quantizing step of a quantizing device 6 is controlled by the size of an action area detected by an action area detector 11, and the size of the action area is detected from the number of the blocks of the area constituted of the block of the periphery having the same action vector as the action vector of an attentional block. At the area of the large action area in a quantizing device 11, many numbers of bits are assigned, and at the area of the small action area, the small number of bits is assigned, and quantized, and thus, the coding is efficiently performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-efficiency coding means used when a moving picture television signal is digitized and transmitted or recorded.

[0002]

2. Description of the Related Art In recent years, information compression means by high-efficiency coding has been used for transmitting or recording a television signal of a moving image, and its compression efficiency has become an important issue.

A conventional high efficiency coding apparatus will be described below with reference to the drawings. Conventionally, in order to compress the amount of information of a television signal of a moving image, means for utilizing correlation between frames has been proposed. In this information compression means, a motion vector between frames of a television signal is calculated. Motion vector detection means for accurate detection is used as an elemental technology. Therefore, first, the motion vector detecting means will be described.

FIG. 6 schematically shows the concept of the motion vector detecting means. In the figure, with respect to the block A (diagonal line) for which the motion vector is to be detected in the current frame, the frame one frame before, the center of which is the same position as the center of the block A is the center of the television signal. The difference between the center of the block and the signals in all the blocks within a certain range R is evaluated, and the difference between the coordinate in the region R giving the minimum value and the center point is detected as a motion vector.

Next, a conventional high efficiency coding apparatus using the above motion vector detecting means will be described with reference to the drawings. FIG. 7 is a block diagram showing the configuration of a conventional general high efficiency coding apparatus. In the figure, 1 is an input terminal of the device, 2 is an output terminal of the device, 3 is a subtractor, 4 is an adder, 5
Is an orthogonal transformer that performs two-dimensional orthogonal transformation, 6 is a quantizer,
Reference numeral 7 is an inverse quantizer, 8 is an inverse orthogonal transformer that performs two-dimensional inverse orthogonal transformation, 9 is a motion vector detector, and 10 is a frame memory for delaying an image by one frame.

The mutual relationship and operation of the above components will be described below. The television signal input from the input terminal 10 of the apparatus is input to the subtractor 3 and the frame memory 1
The difference from the television signal of one frame before stored in 0 is calculated. The orthogonal transformer 5 divides the difference signal between the frames into small blocks and performs two-dimensional orthogonal transformation. When such processing is performed, since the television signal has a high correlation in the horizontal direction and the vertical direction in the time direction, the quantizer 6 allocates many bit lengths in the vicinity of the DC component of the image, and The amount of information is compressed by assigning a small bit length. In this way, the moving image signal with the compressed information amount is output from the output terminal 2 of the apparatus. On the other hand, the inverse quantizer 7, the inverse orthogonal transformer 8, the adder 4, the motion vector detector 9 and the frame memory 10 constitute a circuit for computing the signal one frame before sent to the subtractor 3. First, the television signal compressed by the inverse quantizer 7 is inversely quantized, and then the inverse orthogonal transformer 8 performs two-dimensional inverse orthogonal transformation on the small block performed by the orthogonal transformer 5. .. The adder 4 adds the output from the inverse orthogonal transformer 8 and the output of the frame memory 10 to restore the television signal of the current frame. The motion vector detector 9 detects the motion vector of the television signal by the means shown in FIG. The frame memory 10 uses the motion vector detected by the motion vector detector 9 for the decoded television signal output from the adder 4, and compensates the moving portion of the television signal. The frame is delayed by one frame and sent to the subtractor 3. With the above operation, the moving image signal with the compressed information amount is output from the output terminal 2 of the apparatus.

[0007]

In such a conventional high-efficiency coding apparatus, since the same quantizer is used for the still part and the moving part of the image, the moving part of the image is visually recognized. Despite the deterioration of the characteristics, the same quantization as in the stationary portion is performed, and there is a problem that the compression rate cannot be sufficiently increased. As a means to solve this problem, a configuration has been proposed in which the quantizer is selected using the output from the motion vector detector, but the visual characteristics are controlled by the motion vector for each small block. Although the smaller the size of the part is, the more it deteriorates, the more the amount of information is compressed uniformly in the moving part, and there is a problem that the compression rate cannot be made sufficiently high.

The present invention solves the above problems, and an object of the present invention is to provide an encoding apparatus capable of increasing the compression rate of information.

[0009]

In order to achieve the above object, the present invention detects a motion vector of a moving image signal for each block of an image by a motion vector detecting means and stores the detected motion vector for one frame. A region formed by an arbitrary one block of interest and a peripheral block having the same motion vector as the motion vector of the target block is set as a motion region, and the motion vector of the target block and blocks around the target block Motion vector detecting means for sequentially comparing the motion vectors of the motion vector with the motion vector in a square lattice pattern to detect the size of the motion area by the number of blocks having the same motion vector, and the quantizing means according to the size of the motion area. A high-efficiency coding apparatus for a moving picture signal, comprising: a quantizing means whose quantizing step is controlled.

[0010]

According to the present invention, in the above structure, the moving area detecting means detects the size of the moving area including the block of interest, performs high-quality coding on the stationary portion, and determines the moving area size on the moving portion. The quantization step of the quantizing means is controlled in accordance with the quantization step so that the large part of the moving part area is quantized by the dense quantizing step and the small part of the moving part area is quantized by the coarse quantizing step. To convert.

[0011]

(Embodiment 1) A high efficiency coding apparatus according to an embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a high efficiency coding apparatus according to an embodiment of the present invention. In the figure, 1 is an input terminal of the apparatus, 2 is an output terminal of the apparatus, 3 is a subtractor, 4 is an adder, 5 is an orthogonal transformer for performing two-dimensional orthogonal transformation, 6 is a quantizer, and 7 is an inverse quantum. , 8 is an inverse orthogonal transformer that performs two-dimensional inverse orthogonal transformation,
Reference numeral 9 is a motion vector detector, 10 is a frame memory for delaying an image by one frame, and 11 is a motion area detector for detecting the size of the motion area.

The mutual relationship and operation of the above components will be described below. The television signal input from the input terminal 1 of the apparatus is sent to the subtractor 3 and the difference from the television signal of the previous frame is calculated. The orthogonal transformer 5 divides the difference signal between the frames into small blocks,
Performs two-dimensional orthogonal transformation. When performing such processing,
Since the television signal has a high correlation with the time direction in the horizontal direction and the vertical direction, the electric power is concentrated near the DC component of the image.

Therefore, the quantizer 6 compresses the information amount by allocating a large bit length in the vicinity of the DC component of the orthogonally transformed signal and a small bit length in the high frequency component. In this way, the moving image signal with the compressed information amount is output from the output terminal 2 of the apparatus. On the other hand, the inverse quantizer 7, the inverse orthogonal transformer 8, the adder 4, the motion vector detector 9
And the frame memory 10 constitutes a circuit for calculating the signal one frame before sent to the subtractor 3. In this circuit, the inverse quantizer 7 inversely quantizes the compressed moving image signal, and then the inverse orthogonal transformer 8 two-dimensionally inversely orthogonalizes the small block performed by the orthogonal transformer 5. The conversion is done. The adder 4 adds the output from the inverse orthogonal transformer 8 and the output from the frame memory 10 to restore the television signal of the current frame. The motion vector detector 9 detects the motion vector of the television signal by the means shown in FIG. The frame memory 10 uses the motion vector detected by the motion vector detector 9 for the decoded television signal output from the adder 4 to compensate for the moving portion of the television signal and also to generate one frame. The delay time is sent to the subtractor 3. The output from the motion vector detector 9 is also sent to the motion area detector 11, where the size of the motion area is detected. The quantizer 6 controls the quantization step of the signal of the orthogonal transformer 5 according to the size of the motion area detected by the motion area detector 11.

The motion area detector 11 shown in FIG. 1 will be described below with reference to the drawings. FIG. 2 is a schematic diagram showing the operation of the motion area detector 11. In the figure, the arrow indicates the motion vector of the block, and the arrow direction indicates the direction of the motion vector. The motion area detector detects an area of a block having motion vectors in the same direction. FIG. 3 is a schematic diagram showing the order of searching the motion area. In the figure, the range of blocks to be searched in a square lattice centering on the target block is expanded to the first step and the second step. FIG. 4 is a flow chart showing the operation of the motion area detector. In FIG. 4, in step s1, the motion vector of the target block is set to A, the size (S) of the motion region is set to 0, and the search step (n) is set to 1 as initial values, and the first search step is started. Next, in step s2, the initial value of the number (T) of identical motion vectors in the search range of the nth step of the search is set to 0, but this time, the initial value of the number of vectors (T) of the first step of the search is set. Is set to 0. Then step s
In step 3, the motion vector X of the block within the search range of the nth step of the search is detected and compared with the motion vector A in step s4, but this time it is compared as the first step of the search. When X and A are equal, the process proceeds to step s5, 1 is added to the number (T) of identical motion vectors, and X and A
Is not equal, the process returns to step s3 to move to the processing of the next block. The above process is repeated until the blocks within the search range in the first step of the search run out in step s6. This is the end of the first step of the search. In step s7, the same motion vector number (T) obtained in the first step of the search is checked, and if it is not 0, the same vector number (T) is added to the size (S) of the motion area in step s8. Assuming that the first step is completed, (n), that is, this time, 1 is incremented by 1 to 2 and the process proceeds to step s11. In step 11, the number of steps of the search is compared with the target value N to check,
If (n) = 2 is less than the target value N this time, the process proceeds to step s2 and proceeds to the second step of search. (N)
When = 2 reaches the target value N, the search ends. The above operation is repeated until the Nth step of the search. The search range is expanded until the preset number (N) is reached. If the same motion vector number (T) is 0 in step s7, it is useless to extend the search range further, and the search is ended. By the above processing, when a motion vector for one frame is given, the size of the motion region related to the motion vector of the target block is calculated by sequentially comparing the motion vector in small block units with the surrounding motion vector. (S) is detected.

FIG. 5 is a graph showing the relationship between the quantization step width and the size of the motion area in the quantizer shown in FIG. 1, in which the quantization step width corresponding to the detection result of the motion area detector is shown. Switch. That is, the larger the size of the motion area, the smaller the width of the quantization step and lower the compression rate, and the smaller the size of the motion area, the larger the width of the quantization step and increase the compression rate.

As described above, according to the high-efficiency coding apparatus of the embodiment of the present invention, the motion vector detecting means detects the motion vector of the moving image signal for each block of the image, and the detected motion vector is one frame. A region formed by an arbitrary one block of interest and a peripheral block having the same motion vector as the motion vector of the target block is set as a motion region, and the motion vector of the target block and the periphery of the target block are stored. Motion region detecting means for sequentially comparing the motion vectors of the blocks with the motion vectors of the blocks in a square lattice pattern to detect the size of the motion region based on the number of blocks having the same motion vector, and the quantization according to the size of the motion region. A high-efficiency coding apparatus for moving picture signals, which comprises a quantizing means in which the quantizing step of the means is controlled. The large portion of the allocated a sufficient number of bits like the stationary part, for the small portion of the motion area can be efficiently encoded by quantizing less allocated a number of bits.

[0017]

As is apparent from the above embodiments, the present invention detects the motion vector of the moving image signal for each block of the image by the motion vector detecting means, and stores the detected motion vector for one frame in the storing means. A region composed of an arbitrary one block to be stored and a peripheral block having the same motion vector as the motion vector of the block of interest is defined as a motion area, and the motion vector of the block of interest and the motion of blocks around the block of interest. Motion region detecting means for sequentially comparing vectors in a square lattice pattern to detect the size of the moving region by the number of blocks having the same motion vector; and quantizing means for quantizing means in accordance with the size of the moving region. By providing a high-efficiency coding apparatus for moving picture signals, which comprises a quantizing means whose steps are controlled, efficient moving picture Goka device is obtained.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a high-efficiency coding apparatus according to an embodiment of the present invention.

FIG. 2 is a schematic diagram showing the size of a motion area in the high efficiency coding apparatus according to the embodiment of the present invention.

FIG. 3 is a schematic diagram showing a square lattice search of a motion area detecting means in the high efficiency encoding device according to the embodiment of the present invention.

FIG. 4 is a flowchart showing the operation of the motion area detecting means in the high efficiency encoding apparatus according to the embodiment of the present invention.

FIG. 5 is a graph showing the relationship between the quantization step width of the quantizer and the size of the motion area in the high-efficiency coding apparatus according to the embodiment of the present invention.

FIG. 6 is a schematic diagram showing motion vectors.

FIG. 7 is a block diagram showing the configuration of a conventional high efficiency encoding device.

[Explanation of symbols]

 6 Quantizer (quantization means) 9 Motion vector detector (motion vector detection means) 10 Frame memory (storage means) 11 Motion area detector (motion area detection means)

Claims (2)

[Claims] 1. A motion vector detecting means detects a motion vector of a moving image signal for each block of an image, and the detected motion vector is stored in a storage means for one frame. Is defined as a motion area, and a motion vector of the block of interest and a motion vector of blocks around the block of interest are sequentially compared in a square lattice pattern to obtain the same motion. A method of detecting a motion area, wherein the size of the motion area is detected by the number of blocks having a vector. 2. A motion area detecting means for detecting the size of a motion area of a moving image by the motion area detecting method according to claim 1, and a quantizing means in which a quantization step is controlled by the size of the motion area. A high-efficiency coding apparatus for moving picture signals.