JPH06181568A - Encoding device for video signal - Google Patents

Abstract

PURPOSE:To reduce calculation frequency required for the calculation of motion information in the case of obtaining an encoding device having higher encoding efficiency as compared with a conventional device. CONSTITUTION:A field to be encoded is divided into plural blocks, decoding signals generated before prescribed even fields and odd fields from the field to be encoded are stored and an interpolation signal between respective scanning lines is generated by using the stored signal generated before the odd fields. In the case of detecting the motion information of the encoded block signal from three signals, the signal obtained before the even fields, the signal obtained before the odd fields and the interpolation signal, a motion vector detector 6 detects reference motion information between the signal before the even fields and the encoded block. Motion vector detectors 7, 8 detect the motion information among the odd-field signal, the interpolation signal and the encoded block based upon the reference motion information by narrowing a detection range by means of a divider 9 and a detection range setter 10 and a forecasting signal is generated from the even field signal, the odd field signal and the interpolation signal based upon the motion information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coding apparatus for reducing the data amount of an image and performing high-efficiency coding for recording and transmitting a video signal.

[0002]

2. Description of the Related Art One of video signal coding apparatuses uses motion compensation to reduce redundancy in the time direction. Regarding the processing of the divided blocks, there are one that is performed in frame units and one that is performed in field units. In the current television system, interlaced scanning called interlacing is performed on the screen for each scanning line, and when the screen is not completely stationary, the degree of reduction in redundancy is small even if the difference between frames is taken.

On the other hand, when performing prediction in field units,
There is a drawback in that the position of the scanning line of the block to be encoded is different from that of the prediction signal and the phase is different, so that the prediction may not be performed correctly. In particular, when the screen is stationary, the phase shifts in field units, so it cannot be predicted correctly. For the purpose of efficiently removing temporal redundancy from interlaced images, a processing method for frames, fields, interpolated fields, etc. was previously described in Japanese Patent Application No. 4-207758. We proposed a video signal coding device with high coding efficiency for interlaced images by selecting the most suitable prediction signal for each switching block.

FIG. 3 is a block diagram of a coding apparatus for explaining the coding apparatus. In FIG. 3, 31 is a buffer memory, 32 is a blocker, 33 and 34 are field memories, 35 is an interpolation signal generator, 36, 37 and 38 are motion vector detectors, and 39, 40 and 41 are prediction signal generators. , 42, 43 and 44 are subtractors, 45 is a comparator, 46 is a code amount calculator, 47 is a selector, 48 is an orthogonal transformer, 49
Is a quantizer, 50 is an encoder, 51 is an inverse quantizer, 52
Is an orthogonal transformer, 53 is an adder, and 54 and 55 are switches.

The relationship and operation of these components in the encoding apparatus having the components described above will be described. The image signal for one field stored in the buffer memory 31 is converted by the blocker 32 into vertical 8 pixels ×
Divide into blocks of 8 pixels horizontally. Field memory 3
The decoded signals of one field before and one frame before are stored in 3 and 34, respectively. From the decoded signal one field before stored in the field memory 33, as shown in FIG. 4, a screen corresponding to the position between the scanning lines is displayed by the interpolation signal generator 3
It is generated by 5.

The motion vector detectors 36, 37, 38 use the position corresponding to the position of the block to be encoded for each of the signal of one frame before, the signal of one field before, and the generated interpolation signal as a reference position. The position where the sum of the absolute values of the difference values is the smallest in the detection range is detected as a motion vector, and the block displaced by the motion vector detected from the reference position of each signal is calculated by the prediction signal generators 39, 40, 41. The difference signal with respect to the block to be encoded is calculated by the subtractor 42, 43, 44, which is generated as a prediction signal.

Which of these prediction signals is the most suitable prediction depends on the size and direction of the motion of the block and the definition of the signal within the block. By adaptively selecting this, it becomes possible to sufficiently reduce the redundancy in the time direction even for an interlaced image.

The code amount when the obtained three differential signals are encoded is estimated by the code amount calculator 46, and the magnitude of the obtained three code amounts is compared by the comparator 45. The one having the smallest code amount is selected and encoded by the selector 47. The selected difference signal is orthogonally transformed by the orthogonal transformer 48, quantized by the quantizer 49, and encoded by the encoder 50.

The encoded output is output, the inverse quantizer 51 performs the inverse operation of the quantizer, the orthogonal transformer 52 converts the encoded output again, and the adder 53 adds it to the prediction signal. As a result, it is returned to the video signal.

The returned video signal is stored in the field memory 33.
It is stored in and used for prediction of the next field. The signal of the field memory 33 is transferred to the field memory 34 and used for prediction one frame later.

When a mode for converting and coding the input signal itself of the block to be coded is selected, the switch 54 switches the input to the orthogonal transformer. At this time, the output from the orthogonal converter 48 is the video signal itself and does not need to be added to the prediction signal.

[0012]

However, the above method has a drawback that the number of calculations required for detecting the motion vector increases. Whether the motion vector detector is independent for the three types of motion vector detection, or when one motion vector detector is shared by time-division processing, the motion vectors are calculated independently, so it is necessary for detection. The number of calculations increases.

In view of the above points, an object of the present invention is to provide a video signal coding apparatus which reduces the number of calculations required for motion vector detection and improves coding efficiency for interlaced images as compared with the prior art. .

[0014]

In order to solve the above-mentioned problems, the coding apparatus of the present invention comprises a blocking means for dividing a field to be coded into blocks, and a predetermined value for the field to be coded. An image between scanning lines within a predetermined range by using accumulating means for accumulating decoded signals before even m fields before and n odd fields before a predetermined odd number and signals before n fields accumulated by the accumulating means. Between the scanning line interpolating means for generating a signal by interpolation and the signal of the block to be encoded among the blocks divided by the blocking means and the signal m fields before stored by the storing means. Between the first motion information, the second motion information between the signal n fields before, and the interpolated image signal generated by the scanning line interpolation means. Motion detection means for detecting one type of motion information among the three types of motion signals of the motion information of No. 3 as reference motion information, and the remaining two types of motion information based on the reference motion information detected by the motion detection means. Based on the first motion information detected by the motion detecting means and the motion information calculating means, a first prediction signal and a second motion information from the signal m fields before are calculated. And a prediction signal generating means for respectively generating a second prediction signal from the signal n fields before and a third prediction signal from the interpolated image signal based on third motion information. A first difference signal between the signal of the block and the first predicted signal,
A difference signal calculation unit that calculates a second difference signal between the second prediction signal and the third prediction signal and a first difference signal calculated by the difference signal calculation unit. , The second and third difference signals are compared by a preset method, and the comparison result of the comparison means is used to encode the first difference signal or the second difference signal or It is provided with a selecting means for selecting one of the third differential signal encodings and a means for encoding the differential signal selected by the selecting means.

[0015]

According to the present invention, with the above configuration, it is possible to compare three types of differential signals and switch the processing method between frames, fields, interpolated fields, etc., so that the most suitable prediction signal can be obtained for each block. By selecting, it is possible to obtain a video signal coding device with high coding efficiency for interlaced images, and of course, by utilizing the correlation of three types of motion information, it is possible to calculate three types of motion information. The number of calculations required can be reduced.

[0016]

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 of a video signal encoding apparatus according to an embodiment of the present invention. In FIG. 1, 1 is a buffer memory, 2 is a blocker, 3 and 4 are field memories,
Reference numeral 5 is an interpolation signal generator, 6, 7 and 8 are motion vector detectors, 9 is a divider, 10 is a detection range setting device, 11, 12 and
13 is a prediction signal generator, 14, 15, 16 are subtractors, 1
7 is a comparator, 18 is a code amount calculator, 19 is a selector, 20
Is an orthogonal transformer, 21 is a quantizer, 22 is an encoder, and 23
Is an inverse quantizer, 24 is an orthogonal transformer, 25 is an adder, 2
Reference numerals 6 and 27 are switches.

The relation and operation between these components in the coding apparatus having the above components will be described. The image signal for one field stored in the buffer memory 1 is divided by the blocker 2 into 8 vertical pixels × 8 horizontal pixels.
Divide into blocks of pixels. The field memories 3 and 4 store the decoded signals of one field before and one frame before. From the decoded signal of one field before stored in the field memory 3, as shown in FIG. 2, the interpolation signal generator 5 generates a screen corresponding to the position between the scanning lines.

With respect to the signal one frame before, the position corresponding to the position of the block to be coded is used as a reference position, and the position having the smallest sum of absolute difference values within the detection range is detected by the motion vector detector 6. It is detected as a motion vector of 1. By the divider 9, as shown in FIG.
Vectors that are half the vertical and horizontal sizes of the motion vector are calculated as field reference vectors. Further, the detection range in the inter-field motion detection is set to half of the size defined between the vertical and horizontal frames by the detection range setting unit 10. For each of the signal one field before and the generated interpolated signal, the motion vector detectors 7 and 8 detect the detection range with the position shifted by the size of the field reference vector from the position of the block to be encoded as the reference position. The position having the smallest sum of absolute difference values is detected as the second and third motion vectors. At this time, when detecting the second and third motion vectors, the search start position is shifted based on the magnitude of the first motion vector to limit the detection range. This can reduce the number of calculations required for motion detection. Blocks displaced from the respective signals by the motion vector detected are generated as prediction signals by the prediction signal generators 11, 12, and 13, and the difference signals from the blocks to be encoded are subtracted by the subtracters 14, 15, and 16. calculate.

Whether or not these prediction signals are suitable for prediction varies depending on the size and direction of the motion of the block, the definition of the signal within the block, and the like. By adaptively selecting this, it becomes possible to sufficiently reduce the redundancy in the time direction even for an interlaced image.

A code amount calculator 18 estimates a code amount when the obtained three differential signals are encoded, and a comparator 17 compares the magnitudes of the obtained three code amounts. The selector 19 selects and codes the code having the smallest code amount. The selected difference signal is orthogonally transformed by the orthogonal transformer 20, quantized by the quantizer 21, and encoded by the encoder 22.

The encoded output is output, the inverse quantizer 23 performs the inverse operation of the quantizer, the orthogonal transformer 24 transforms the differential signal again, and the adder 25 adds it to the prediction signal. As a result, it is returned to the video signal.

The returned video signal is stored in the field memory 3 and used for prediction of the next field. The signal of the field memory 3 is transferred to the field memory 4 and used for prediction after one frame.

When a mode for converting and coding the input signal itself of the block to be coded is selected, the switch 26 switches the input to the orthogonal transformer. At this time, the output from the orthogonal transformer 20 is the video signal itself and does not need to be added to the prediction signal.

In the above description, the size of the divided blocks is 8 × 8 pixels, and the size for comparing the differences between frames is 8 × 8 pixels, but these sizes are arbitrary. In addition, the first to second and third motion vector detection
The number of calculations is limited by using the result of the motion vector detection of the second motion vector detection, which is used for the first and second motion vector detection, which uses the result of the second motion vector detection for the first and third motion vector detection. Alternatively, a method using the result of the third motion vector detection may be used. Although the field reference vector is set to half the magnitude of the first vector in the second and third motion vectors, this ratio is arbitrary. Also, the first
The detection range is set to half of the size between frames by the motion vector of, but this ratio is arbitrary. Moreover, although the sum of absolute values of the difference values is compared with respect to all the positions within the detection range, weighting may be performed with the reference position as the center, or a tree search method of several stages may be used. In addition, in motion vector detection, the sum of absolute values of difference values is compared and the smallest one is selected, but other functions such as the sum of squares of difference values may be used. Further, as the comparison method, a method of preparing one or a plurality of threshold values and selecting one from the respective judgments may be used.

Although the predicted screen is one frame before and one field before, the screens before one frame and one field or more may be used as long as they are selected according to the phase of the interlace. Also, the number of prediction screens and the number of interpolation screens generated by interpolation may be increased. Further, although the orthogonal transform is used in the encoding process, a vector quantization method or the like may be used regardless of the processing method and the quantization method.

Further, although the description has been given as the configuration diagram having the motion detector, the prediction signal generator, and the subtractor for each prediction signal, the same one can be used by the time division processing. The code amount calculator is configured to share the same one, but it may be configured to have it for each prediction signal. In addition, it was said that the one that actually encodes is selected from the three types of difference signals by calculating and comparing the code amount for each prediction signal. This is a simpler comparison criterion such as sum of absolute values or sum of squares. May be used.

[0027]

As described above, according to the present invention,
It is possible to reduce the number of calculations required for motion information calculation, and it is possible to obtain a video signal coding apparatus with higher coding efficiency for interlaced images.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a prediction method and an interpolation method for explaining an encoding device according to an embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a conventional encoding device.

FIG. 4 is a prediction method for explaining a conventional encoding device,
Diagram showing the interpolation method

[Explanation of symbols]

 2 Blocker 3, 4 Field memory 5 Interpolation signal generator 6, 7, 8 Motion vector detector 9 Divider 10 Detection range setting device 11, 12, 13 Prediction signal generator 14, 15, 16 Subtractor 17 Comparator 18 code amount calculator 19 selector 22 encoder

Claims (5)

[Claims] 1. A blocking means for dividing a field to be encoded into blocks, and a decoded image signal before a predetermined even m field and a predetermined odd n field before the field to be encoded, respectively. Storing means, a scanning line interpolating means for generating an image signal between scanning lines within a predetermined range by interpolation using the decoded image signal n fields before stored by the storing means, and the blocking means. For the signal of the block to be encoded among the blocks divided by, the decoded image signal before the m field or the decoded image signal before the n field stored by the storage means or generated by the scanning line interpolation means Motion information is detected as reference motion information from one of the three types of interpolated image signals. Based on the motion detecting means and the reference motion information detected by the motion detecting means, two kinds of motion information of the remaining two kinds of image signals accumulated in the accumulating means and the block to be encoded are calculated. A first prediction signal from the m-field-preceding signal based on the first motion information between the motion-information calculating means for calculating and the m-field-preceding signal detected by the motion detecting means and the motion-information calculating means. , And a second prediction signal based on the second motion information between the n-field-preceding signal and the n-field-preceding signal, and an interpolating image signal generated by the scanning line interpolating means. Prediction signal generating means for respectively generating a third prediction signal from the interpolated image signal based on the motion information of No. 3, the signal of the block to be encoded, and the first prediction signal. First difference signal between the second difference signal between the second prediction signal, the third
Difference signal calculating means for calculating a third difference signal of the predicted signal and the first, second, and third difference signals calculated by the difference signal calculating means are compared by a preset method. Selecting means for selecting one of the encoding of the first difference signal, the encoding of the second difference signal, and the encoding of the third difference signal from the comparison means and the comparison result of the comparison means. And a video signal coding apparatus, characterized in that it comprises means for coding the differential signal selected by the selecting means. 2. The video signal encoding apparatus according to claim 1, wherein the remaining two types of motion detection ranges are limited based on the reference motion information. 3. The method according to claim 1, wherein the remaining two types of motion detection reference positions are moved based on the reference motion information.
Alternatively, the video signal encoding device described in 2. 4. The video signal coding apparatus according to claim 1, wherein the even number m field and the odd number n field have a relationship of m = n + 1. 5. A code amount calculating means for obtaining a code amount when encoding the first, second and third differential signals, respectively,
4. The video signal encoding device according to claim 1, wherein the code amounts obtained by the code amount calculating means are compared with each other.