JPH0670310A - Moving picture discrimination device for coding television picture - Google Patents

Abstract

PURPOSE:To provide the device to discriminate a moving picture for efficiently coding a television picture signal when the signal is subjected to high frequency coding through the use of discrete cosine transformation. CONSTITUTION:The device is provided with a shift register 2 separating a television picture signal subjected to block processing in horizontal and vertical directions in a frame memory 1 into a same field and a different field, a same field correlation calculation circuit 3 calculating correlation between picture elements in the same field in a block, a different field correlation calculation circuit 4 calculating correlation between picture elements in the different field in a block, and a moving picture discrimination circuit 5 comparing outputs of both the circuits 3,4 to discriminate the moving picture.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving image discriminating apparatus for television images used in the field of video technology.

[0002]

2. Description of the Related Art With the digitization of image signals, high efficiency coding technology has become important. Therefore, peripheral technology for efficiently encoding the image signal is also required. For example, when encoding a television image signal, it is possible to perform more optimal high-efficiency encoding by determining whether the television image signal is a signal having a property such as a moving image or a still image.

In the conventional moving image discriminating apparatus, one frame of a television image signal is divided into blocks of about 8 × 8 pixels,
The sum of the pixel values in each block is calculated and compared with the pixel values of the other blocks to detect the motion vector of the image, or the difference between the blocks of the TV image signal frame is calculated, and the difference value is calculated. There is a method to detect the direction of movement, the distance moved, etc. from a block with a certain value or more.

[0004]

However, a method of detecting a motion vector by comparing the sum of the pixel values in the above-mentioned blocks, a direction of the motion of the block due to a difference between blocks of a frame of a television image signal, The method of detecting the distance increases the line scale. In addition, when performing high-efficiency coding based on the result of image discrimination, especially when encoding using discrete cosine transform (hereinafter abbreviated as DCT), what is determined as a moving image is hand,
It is not clear if it can be used effectively.

The present invention has been made in view of the above conventional problems, and an object of the present invention is to provide an efficient moving picture discriminating device for a television image, particularly for high efficiency coding using DCT.

[0006]

According to the present invention, a block forming means for dividing a television image signal converted into a digital signal into blocks of H pixels in the horizontal direction and V pixels in the vertical direction in one frame. And for each pixel of the television image signal blocked by the blocking means, the same-field correlation value obtained from pixels in adjacent rows in the same field and different-field correlation obtained from pixels in adjacent rows in different fields. The values are used to perform weighting, and a means for judging the moving image is provided based on the degree of the weighted values in the entire block.

[0007]

According to the present invention, a television image signal converted into a digital signal is blocked in a frame, and the pixels of the blocked television image signal are the same between pixels in adjacent rows in the same field. The field correlation value and the different field correlation value between adjacent rows of pixels between different fields are used for weighting to find out whether the intra-field correlation or intra-frame correlation is strong, It is determined whether the image inside has a property such as a moving image or a still image.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS 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 an embodiment of the present invention. In FIG. 1, 1 is a frame memory, 2
Is a shift register, 3 is the same field correlation value calculation circuit,
Reference numeral 4 is a different field correlation value calculation circuit, and 5 is a moving image determination circuit.

FIG. 2 shows that the input television image signal of the embodiment shown in FIG. 1 is divided into H pixels in the horizontal direction and V pixels in the vertical direction.
FIG. 4 is an explanatory diagram of a block having individual pieces, and FIG. 3 is an explanatory diagram showing a difference in field of an input television image signal. Since the television performs interlaced scanning, one frame is composed of the first field and the second field. That is, the odd rows A, C, B, ... Of FIG. 2 are the first field, the even rows B, D ,.
... belong to the second field, respectively. Therefore, a row adjacent to any C row is A if it is in the same field.
Line, E line, B line between different fields,
It is row D.

Hereinafter, the same field and different fields are in the same frame, and are divided and H is set to 8 in the horizontal direction.
A case of exemplifying FIG. 2 in the case of 16 blocks of V in the vertical direction will be described.

A television image signal is input to the frame memory 1 and recorded for one frame. This recorded 1-frame data is divided into 8 blocks in the horizontal direction and 16 blocks in the vertical direction to output 8 × 16 data blocks, which are output for each block. The horizontal row data of the blocked signals are input to the shift register 2 and distributed to the signals for each field in the shift register, and the same field correlation value calculation circuit 3 and the different field correlation value calculation circuit 4 are provided. Are input to the circuit, the respective correlation values are calculated, and the moving image determination circuit 5 determines the moving image.

Next, the process from the blocked signal to the image determination will be described in detail with reference to FIG.

FIG. 4 is a block diagram showing a detailed configuration excluding the frame memory of FIG. In the figure, 6 to 9 are shift registers of 8 clock delay (8L), which constitute the shift register 2 of FIG. 10, 11 are selectors, 12,
13, 20, and 21 are adders, 14 and 15 are dividers that reduce the number to half,
Reference numerals 16, 17 and 26 are subtractors, 18 and 19 are absolute value circuits, 22 and 23 are line-by-line reset circuits, 24 and 25 are latch circuits, 27 and 29 are comparators, and 28 is a counter.

The output from the frame memory 1 is divided into blocks, and the signals of each row are divided into shift registers 2 (see FIG. 4).
Shift register 6, shift register 7, shift register 8, shift register 9), and signals En, D from connection points e, d, c, b, a of each shift register in FIG.
n, Cn, Bn, An are output (n is 1 to 8),
Sequentially move to the next line. Here, En, Cn, and An belong to the first field of the same field, and Dn and Bn belong to the second field of different fields, and the fields are separated. For example, the inputs of the selector 10 are A1 and E shown in FIG.
1, the outputs are A1 and E1, and the input of the selector 11 is shown in FIG.
The outputs are also B1 and D1. In this case, the average value of A1 and E1 is output by the adder 12 and the divider 14, and the predicted value from the same field is obtained. Further, the average value of B1 and D1 is output by the adder 13 and the divider 15, and the predicted value from different fields is obtained.

Next, the absolute value of the difference between each average value and C1 is obtained by the subtracters 16 and 17 and the absolute value circuits 18 and 19. In terms of fields, as shown in FIG. 3, the correlation between C and A, E in the same field,
It shows the correlation between different fields in which C is compared with B and D.

Next, the correlation value for each pixel is reset for each row by the reset circuits 22, 23 and added by the adders 20, 21, and the adders 20, 21 and the reset circuits 22, 23 and the latch circuit are added.
Calculated by 24, 25. The latch circuit 24 obtains the sum of correlation values (Equation 1) for eight pixels in the same field, and the latch circuit 25 obtains the sum of correlation values (Equation 2) for eight pixels in different fields. Subtract this difference
26, and the comparator 27 compares them as the numerical value given by the first designated value M. For example, when (Equation 2) is larger than (Equation 1) by M or more, the enable pulse is output from the comparator 27. It

[0018]

[Equation 1]

[0019]

[Equation 2]

As a result, the counter 28 is counted up, the counter 28 is reset for each 8 × 16 block, and finally the counter value is compared with the numerical value given by the second designated value N by the comparator 29 to obtain N. If it is larger than the above, it is determined to be a moving image.

That is, the correlation for each pixel is summed up for each line, and the number of lines in which the difference in the correlation value for each line between the same field and a different field exceeds the first designated value M is the second designated value of a fixed number of lines. When N is exceeded, the moving image is judged. This allows a simple circuit to determine the moving image for each 8 × 16 block, and to know which of the intra-field correlation and the intra-frame correlation the pixels in the 8 × 16 block have. For example, when performing a calculation using DCT after this,
It is easier to determine which of the field-based DCT and the frame-based DCT contains the low frequency component. For example, if the calculation such that the DCT is switched in the field unit or the frame unit according to the result of this determination is performed, more efficient encoding can be performed.

In order to calculate the correlation value in the 8 × 16 block more accurately, the correlation value at the line toward the end of the block is, for example, when the output of the shift register 7 in FIG. 4 is An. Selects both outputs of the selector 10 to Cn, outputs both of the selector 11 to Bn, and performs subsequent calculations. When the output of the shift register 7 is Pn, sets both outputs of the selector 10 to Nn. If both outputs of 11 are turned on and the subsequent calculation is performed, the correlation value calculation completed within the block can be performed.

[0023]

As described above, in the moving picture discriminating apparatus for television images of the present invention, the pixel in the divided block is compared with the pixel in the same field and the correlation between the pixels in different fields. By doing so, it is possible to efficiently determine a moving image and a still image in high-efficiency encoding using DCT.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

2 shows H input TV image signals in FIG. 1 in the horizontal direction,
It is an explanatory view of a block divided into V pieces in the vertical direction.

FIG. 3 is an explanatory diagram showing a difference between fields of an input television image signal.

4 is a block diagram showing a detailed configuration excluding the frame memory of FIG. 1. FIG.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Frame memory, 2 ... Shift register, 3 ... Same field correlation value calculation circuit, 4 ... Different field correlation value calculation circuit, 5 ... Movie determination circuit, 6-9 ... Shift register, 10, 11 ... Selector, 12, 13, 20, 21 ... Adder,
14,15 ... Divider to reduce to 1/2, 16,17,26 ... Subtractor, 18,19 ... Absolute value circuit, 22,23 ... Reset circuit,
24, 25 ... Latch circuit, 27, 29 ... Comparator, 28 ... Counter, a, b, c, d, e ... Connection point, M ... First designated value, N ... Second designated value.

Claims (2)

[Claims] 1. A block forming means for dividing a television image signal converted into a digital signal into blocks of H pixels in the horizontal direction and V pixels in the vertical direction in one frame, and a block by the block forming means. For each pixel of the converted TV image signal, weighting is performed using the same-field correlation value obtained from pixels in adjacent rows in the same field and the different-field correlation value obtained from pixels in adjacent rows in different fields. And a moving image discriminating apparatus for a television image, comprising means for discriminating a moving image according to the degree of the weighted value in the entire block. 2. A horizontal total of absolute values of differences between a pixel in an arbitrary row and a predicted value obtained from adjacent upper, lower, or upper and lower pixels in the same field, and a pixel in the row. The absolute value of the difference between the predicted value obtained from the upper, lower, or upper and lower pixels between different fields is compared with the sum of H horizontal values, and the difference value of the compared values is the first specified value or more. 2. The moving picture discriminating apparatus for a television image according to claim 1, further comprising means for discriminating a moving picture based on a moving picture and a moving picture having a second specified value or more among V pieces in the vertical direction.