JP3469626B2 - Motion compensated video signal processing apparatus and video signal processing method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion compensation video signal processing system.

BACKGROUND OF THE INVENTION The processing of motion-compensated video signals is used in applications such as television format conversion, film format conversion and format conversion between video and film.

In motion-compensated television converters, such as the converter disclosed in British Patent Application No. GB-A-2231749, a series of input image pairs are processed and the image between the pairs of input images is processed. Generate a set of motion vectors that represent motion. This process is performed individually for each image block so that each motion vector represents the motion between images of the contents of the respective block.

Each set of motion vectors is provided to a motion vector reducer to produce a subset of the set of motion vectors for each block. This subset is sent to a motion vector selector, where it assigns one of the motion vector subsets to a pixel in each block of the image. The selected motion vector assigned to each pixel is sent to a motion compensated interpolator, which interpolates the output image from the input image and takes into account motion between the input images.

[0004]

In this case, when the input image consists of a series of fields of an interlaced video signal, if the image moves vertically by an odd number of video lines between consecutive fields, the vertical Information is lost. Further, this causes a noise-like discrepancy (which causes subjective confusion) in the vertical image movement in the odd and even lines. This point will be described later.

Therefore, an object of the present invention is to smoothly compensate for the loss of image information that becomes noticeable when there is vertical image movement, especially image movement between odd-numbered and even-numbered lines, and to smoothly transform the image. Motion compensation that can
A video signal processing device and a video signal processing method are provided.

[0006]

The present invention provides input
Generate a motion vector that represents the motion of the image between the image pairs,
The motion that forms the output image by motion-compensated interpolation
Compensation video signal processing device, when there is a motion of the image
For each input image pair, half-bandwidth interpolation processing is performed.
To produce two intermediate pixel values, then
Two intermediate pixel values are time-weighted, combined, and output
The means to generate the pixels of the field and the movement of the image
If not, interleave the input image pairs, then
Full bandwidth interpolation processing for interleaved pixels
To generate the pixels in the output field.
A motion-compensated video signal processing device comprising:
It

According to the invention, means are provided for detecting whether each pixel of the output image represents a dynamic part or a static part of the output image. This information is used, for example, in the adaptation of the interpolation method that produces this pixel, or in the subsequent down conversion of the output image.

In order to communicate the detection result as to whether the pixels of the output image are in the dynamic part or the static part to the other parts of the device, the device sets the motion flag associated with each pixel of the output image. Means for generating are preferably provided and this flag indicates whether this pixel was detected to be in the dynamic or in the static part of the output image.

In order to reduce the complexity of the device, it is desirable to detect the output pixel motion as part of the motion vector selection process. To this end, the apparatus comprises means for generating a plurality of motion vectors, including a zero motion vector, for each pixel of the output image, and testing the motion vectors to select a motion vector to be used for interpolating the pixels of the output image. Means comprising means for detecting a degree of correlation between blocks under test of an input image pointed to by a motion vector under test, and blocks under test pointed to by the motion vector from the plurality of motion vectors Means for selecting the motion vector having the highest degree of correlation.

The preferred embodiment apparatus comprises a motion-compensated interpolator that produces each pixel of the output image from each pair of input images according to the motion vector selected for use in interpolating that pixel.

Detection of output pixel motion can be used to alter the behavior of the interpolator. In particular, the motion-compensated interpolator comprises two pixel interpolators and means for selecting one of the two pixel interpolators, depending on whether the pixel is in the dynamic part of the output image or not. It is desirable to interpolate the pixels of the output image according to the detection result of whether there is any.

For example, if the input image consists of a series of fields of interlaced video signals, vertical motion of the picture by an odd number of video lines between the series of fields results in loss of vertical information. This situation can be detected by detecting the odd vertical component of the motion vector selected for interpolation of the output pixel. However, this causes a noise-like (or subjectively disturbing) discrepancy (difference) in the vertical movement of the image between the odd-numbered line and the even-numbered line. Therefore, it is desirable to distinguish between dynamic pixels and non-dynamic pixels. in this case,
In case no image motion is detected, one of the pixel interpolators is interleaved with the means for interleaving the pixels from the two input images to produce a block of interleaved pixel values. And means for producing pixels of the output image by intra-block interpolation of blocks of pixel values.
Similarly, in case an image motion is detected, one of the pixel interpolators has means for generating two respective intermediate pixel values by intra-image interpolation of each pair of input images, and the two And means for combining the intermediate pixel values to produce the pixels of the output image.

The combining means operates to combine the two intermediate pixel values according to a combination ratio depending on the temporal position of the output image with respect to the pair of input images. In one preferred embodiment, a linear relationship is used so that the combining ratio is proportional to the position in time of the output image with respect to the pair of input images.

Preferably, the apparatus operates on a pair of input images consisting of a series of fields of an interlaced input video signal, and the output image consists of a field of interlaced output video signals. By using the interlaced input image, the need for progressive scan conversion of the input video signal is eliminated and the required processing is significantly reduced.

According to a second aspect of the present invention, between input image pairs
Motion vector that represents the motion of the image of
Motion compensated video forming the output image by interpolation
A signal processing apparatus, a plurality of motion compensation pixel interpolator, selecting one from the plurality of motion compensation pixel interpolator in accordance with the motion detection of the image between the input image pair from the input image pair And a means for performing pixel interpolation of the output image.

The device according to the invention is particularly advantageous for use in a television conversion device.

A third aspect of the present invention is a video signal processing method for generating a motion vector representing the motion of an image between a pair of input images in order to generate an output image by motion compensation interpolation. If there is motion in the image,
Half-bandwidth interpolation processing is applied to each of the two intermediate peaks.
Generate the xel value and then the intermediate pixel between these two
Pixels in the output field are time-weighted and combined
Input image when there is no image movement.
Interleaved pairs, then interleaved
By performing a full bandwidth interpolation process on the pixel
Generating a pixel of the output field of the video signal .

According to a fourth aspect of the present invention, there is provided a motion-compensated video signal processing method for forming an output image by motion-compensated interpolation according to a result of detection of image motion between a pair of input images. , A method of selecting one from a plurality of pixel interpolators and interpolating a pixel of an output image from a pair of input images according to a detection result of image motion between the pair of input images.

Embodiments of the present invention will be described by way of example only with reference to the accompanying drawings, wherein like parts are designated with like reference numerals.

[0020]

1 is a block diagram showing an outline of a motion compensation television system conversion apparatus. This device is provided with an input of interlaced digital video signal 50 (eg 112
5/60, 2: 1 high definition video signal (HDVS)) and produces an output of interlaced digital video signal 60 (eg 1250/50, 2: 1, HDVS).

The input video signal 50 is first provided to the input buffer / packer 110. For normal definition input signals, this input buffer / packer 110 converts the input image into a high definition (16: 9 aspect ratio) format and fills black pixels where needed. For HDVS input, the input buffer / packer 110 only provides temporary storage of data.

The data is sent from the input buffer / packer 110 to a matrix circuit 120 where the format of the input video signal (if required) is standard "CCI".
R recommendation 601 ”(Y, Cr, Cb) format. The input video signal from the matrix circuit 120 is transmitted to a time base conversion (TBC) delay device 130 or is transmitted via a subsampler 170 to a sub (down) sampled time base conversion delay device 180.
The time-base conversion delay device 130 determines the temporal position of each field of the output video signal, selects the two fields of the input video signal that are closest in time to the output field, and uses them for interpolation of this output field. For each field of the output video signal, the two input fields selected by the time base converter are given an appropriate delay before being fed to the interpolator 140, where the output field is interpolated. The control signal t indicating the time position of each output field with respect to the two selected input fields is output from the time base conversion delay unit 130 to the interpolator 1.
Sent to 40.

Subsample time base conversion delay device 1
80 operates in a similar manner, but uses the spatially subsampled video signal provided by subsampler 170. A pair of input fields is selected from the subsampled video signal by a subsample timebase conversion delay 180 and used to generate a motion vector.

The timebase converters 130 and 180 operate according to a sync signal associated with the input video signal or the output video signal or both. If only one sync signal is provided, the timing of the other field of the two video signals is based on the time base converters 130 and 18.
Generated as deterministic within 0.

Subsample time base conversion delay unit 18
The pair of fields of the sub-sampled input video signal selected by 0 from the direct block matcher 190, data stripper 200, motion vector estimator 210, motion vector reducer 220, motion vector selector 230 and vector post-processor 240. To the motion processor 185. The pair of input fields is first directly supplied to the block matcher 190, where the search block of the two selected input fields that is ahead in time and the pair of the two input fields that are behind in time (more than big)
Calculate a correlation surface that represents the spatial correlation with the search area. The data representing these correlation surfaces is reformatted by the data stripper 200 and sent to the motion vector estimator. The motion vector estimator 210 detects a point in the correlation surface where the correlation is maximum. (The correlation surface actually represents the difference value between blocks of two input fields. This means that the point at which the correlation is maximum actually takes the minimum value on the correlation surface. The "minimum value".) In order to detect one minimum value, the loss of resolution caused by adding and interpolating points on the correlation surface and using the sub-sampled video signal for generating the surface, Compensate to some extent. From the detected minimum value of each correlation surface,
The motion vector estimator 210 causes a motion vector to be generated,
This is supplied to the motion vector reducer 220.

The motion vector estimator 210 performs a confirmation test on each generated motion vector to confirm whether or not the motion vector is considerably higher than a general noise level, and the result of this confirmation test is obtained. The confirmation flag shown is associated with each motion vector. This confirmatory test, called the threshold test, is described in British patent application GB-A-22.
No. 31749 (along with the features of the apparatus of FIG. 1).

Further testing is performed by the motion vector estimator 210 to detect if each vector is authentic. In this test, the correlation surface (excluding the excluded region around the detected minimum value) is examined to detect the next minimum value. If this second minimum does not exist at the boundary of the exclusion region, then the motion vector derived from the original minimum is flagged as a potential false vector.

The motion vector reducer 220 performs the operation of reducing the possible branches of the motion vector for each pixel of the output field and subsequently supplies the motion vector to the motion vector selector 230. The output field is virtually divided into blocks of pixels, each block being provided with a position in the output field that corresponds to the position of the search block earlier in the selected input field. . The motion vector reducer organizes a group of four motion vectors, associates each block of the output field with each pixel in this block,
Final interpolation is performed using one selected from the group of four motion vectors.

Vectors flagged as "false vectors" are requalified in the vector reduction process when they are identical to the unflagged vectors of adjacent blocks.

As part of this function, motion vector reducer 220 requalifies "qualified" motion vectors (ie, motion vectors that have passed confirmation and false positive tests, or no false positives). Generated motion vectors) are counted and the block positions of the input fields used to obtain these motion vectors are not taken into account. Eligible motion vectors are ranked in descending order of frequency. The most common vector among the eligible motion vectors that are clearly different from each other is classified as a "global" motion vector. The three motion vectors that pass the verification test are selected for each block of output pixels and provided to the motion vector selector 320 with zero motion vectors for further processing. These three selected motion vectors are selected according to the following predetermined priority order. (1) Motion vector generated from corresponding search block (2) Motion vector generated from surrounding search block (local motion vector) (3) Global motion vector

The motion vector selector 230 receives as inputs the two input fields selected by the subsample time base transform (TBC) delay unit 180 and used to derive the motion vector. These fields are given an appropriate delay so that they are supplied to the motion vector selector 230 at the same time as the vectors generated from these fields. Motion vector selector 230 provides an output to which one motion vector per pixel in the output field corresponds. This motion vector is selected from among the four motion vectors provided by the motion vector reducer 220 for this block.

The vector selection process (discussed in more detail below) involves detecting the degree of correlation between the blocks under test of the two input fields identified by the motion vector under test. The motion vector with the highest degree of correlation between test blocks is selected and used to interpolate the output pixels. A “motion flag” is also generated by the vector selector 230. If the degree of correlation between blocks identified by a motion vector of zero is greater than a predetermined threshold, then this flag is set as "quiet" (no motion).

The vector post-processor 240 reformats the motion vector selected by the motion vector selector 230 so that the vertical scale of the image is reflected, and the reformatted vector is interpolated by the interpolator 1.
Supply to 40. By using the motion vector, the interpolator 140 interpolates and produces an output field from the corresponding two (non-subsampled) interpolated input fields selected by the time base conversion delay 130. It takes into account the motion of the image represented by the motion vector supplied to the interpolator 140.

When the motion flag indicates that the current output pixel is in the moving part of the image, the pixels from the two selected fields supplied to this interpolator are:
The two input signals (denoted by the control signal t) are combined at a relative ratio depending on the temporal position of the output field, with a larger ratio being used for the input field closer to the output field. What the motion flag indicates is "quick stop"
If it is set to, the weight distribution over time is not performed.
The output of the interpolator 140 is supplied to the output buffer 150 and is output as the HDVS output signal, and is also supplied to the down converter 160, where the motion flag is used to generate the output signal 165 of normal definition.

The subsampler 170 is a matrix circuit 1
Perform horizontal and vertical spatial subsampling of the input field received from 20. In this horizontal sub-sampling, the input field is first pre-filtered by a half band low pass filter (2: 1 horizontal decimation in this example) and then every other video along each video line. Samples are discarded, cutting the number of samples on each video line in half.

Vertical subsampling of the input field is complicated because the input video signal is interlaced. This has the same effect as the series of lines of video samples in each interlaced field being separated by two video lines in the completed frame, and each field line in the completed frame's video line. Vertical shift from the line of the field immediately before or immediately after. The actual method of vertical sampling used is the first step of vertical low pass filtering (to avoid aliasing), followed by each pixel moving vertically down half a video line ( Filtering operation in case of even field) or moving upward (in case of odd field). The resulting field is roughly equivalent to a vertically subsampled progressive scan frame by a factor of two.

FIG. 2 is a diagram showing an outline of the operation of the motion vector selector 230.

As described above, the motion vector selector 23
0 receives the local and global motion vectors from the motion vector reducer 220, the two subsampled input fields that generate the motion vectors, and the control signal t from the timebase conversion delay 180. For each pixel 300 of the current output field 310, the motion vector selector 230 compares the block 340 of pixels pointed to by each of the preceding and succeeding input fields 320, 330 motion vectors,
Test each of the four possible motion vectors for that pixel.

The comparison operation of the block under test 340 is performed by calculating the sum of the absolute values of the brightness differences between the corresponding pixels in the two blocks, and the smaller the sum, the higher the correlation between the blocks. There will be. This comparison operation is performed on each of the four motion vectors supplied to the motion vector selector 230, but for clarity of the drawing, in FIG. Motion vector 350 and another non-zero motion vector 360) are shown.

FIG. 3 is a schematic block diagram of the motion vector selector 230. The four motion vectors for each block of output pixels are provided by motion vector reducer 220 to motion vector selector 230. These four motion vectors, namely the zero motion vector and v1, v
The three other vectors denoted by 2, v3 are provided to each of the four processing units 400, 410, 420 and 430. Processing units 440, 410, 420, 4
Each of the 30 includes two random access memories (RAMs) 450, 460 and blocks each storing an associated portion of each of the pairs of input fields selected by the address offset calculator 440, the time base conversion delay unit 180. A butt comparator 470 is provided.

In each processing unit, the address offset calculator 440, together with the time offset control signal (t) generated by the time base conversion delay unit 180,
The motion vector assigned to this processing unit is received. From these two values, the Address Offset Calculator will access the RAM 450, to access the block under test of the pixel pointed to by the motion vector.
Generate a plurality of memory addresses to be stored at 460. In response to the address provided by the address offset calculator, each RAM 450, 460 provides an array of pixel values representing a test block to a block match comparator 470.

The block matching comparator 470 compares pixel values (especially luminance components of pixel values) at corresponding positions of two test blocks. The absolute value of the luminance difference value between each pixel pair is calculated and a sum of absolute luminance difference values (SAD) is generated to represent the overall correlation between the two test blocks. SA for a specific motion vector
When the D value is low, it indicates that the degree of correlation between the test blocks pointed to by this motion vector is high.

The processing unit 430 receives the motion vector v3 and calculates the SAD value for the block under test pointed to by this motion vector. Processing unit 4
30 then supplies the SAD value and the identifier of the vector v3 to the processing unit 420.

In the processing unit 420, the vector v
SAD values for 2 are calculated by the respective block matching comparators 4
It is generated from 70. The SAD value is then compared to the SAD value for v3 received from processing unit 430. When this SAD value is low, the motion vector (v2 or v3) indicates that the degree of correlation between the test blocks pointed to by this motion vector is high. Therefore, the processing unit 420 outputs a vector identifier 480 that identifies the lower SAD value 475 and the vector in which the lower SAD value was generated.

The processing unit 410 receives the vector v1 and calculates the SAD value from this vector. The SAD value for v1 is compared to the lower SAD value 475 of vectors v2 and v3 supplied from processing unit 420. The processing unit 410 now has vectors v1, v
The lowest SAD value 477 of 2 and v3 is the SA
The D value is output together with the generated vector identifier 485.

Processing unit 400 generates a SAD value for the zero motion vector, which is processed by processing unit 410.
Compare with the current minimum SAD value for the vectors v1, v2, v3 received from By this comparison process, 4
The selected vector identifier 490 representing the one of the vectors that produces the smallest SAD value to the processing unit 4
It is output from the block matching comparator 470 in 00.

The SAD value for the zero motion vector (generated by block match comparator 470 in processing unit 400) is provided to comparator 500 where it is compared to a preset threshold value 510. The SAD value for the zero motion vector is four vectors (zero, v
1, v2, v3), regardless of which one was selected for use in interpolating the output pixel 300, is provided to the comparator 500.

Comparator 500 sets motion flag 520 (indicating that there is motion in the image) when the SAD value for the zero motion vector is greater than threshold 510. When the SAD value for the zero motion vector is less than the threshold value 510, the motion flag is not set, indicating that the current output pixel is substantially in the stationary portion of the image.

As mentioned above, the operation of interpolator 140 changes depending on whether the motion flag is set for the output pixel being interpolated. The reason for this change is shown in Figure 4.
The description will be made with reference to FIGS.

FIG. 4 is a schematic diagram showing a portion of a progressively scanned video frame of a diamond. When an image is shown as two consecutive interlaced video fields, two video fields are displayed, as shown in Figure 5a (odd video field) and Figure 5b (next even video field). Are recombined to produce a complete diamond pattern, so there is no vertical information loss.

However, information is lost in the interpolation process when there is only vertical odd-numbered vertical image motion between two consecutive video fields. This state is shown in a and b of FIG. FIG. 6a represents an odd video field derived from the pattern shown in FIG. 4, and FIG.
It represents the next (even) video field after the image has moved vertically by one line. The diagrams shown in Figures 6a and 6b are identical, in other words in this case half of the image information is lost.

To solve the above problems, the interpolator 14
0 switches between the two operating modes depending on whether the motion flag is set. When the motion flag is set, it indicates that the current output pixel forms a moving part of the image and intra-field interpolation of each of the pairs of input fields produces two intermediate pixel values. . These intermediate values are combined to produce the output pixel. In contrast, when the motion flag is not set, it indicates that the current output pixel does not form a moving part of the image, and the pixels from the two input fields are interleaved and this interleaved A single filtering operation is performed on the selected pixels to generate output pixels.

FIG. 7 is a diagram showing the operation of the interpolator when the motion flag is set. Half-bandwid
th) Interpolation filter 600 has two input fields 32
Respectively applied to 0, 330 to produce respective intermediate pixel values 610, 620. Half-bandwidth filters are used to avoid the vertical aliasing problem.

Here, two intermediate pixel values 610,
620 are combined according to the time weighting that the field closest in time to the current output field is used in large proportion to form the intermediate pixel to produce the output pixel. In fact, the coupling ratio varies linearly according to the temporal position (t) of the output field between the two input fields. If t is a value between 0 and 1, this indicates that the temporal distance between the output field and the preceding input field is part of one input field period, so two The intermediate pixels are combined as follows. Output pixel = (1-t) × (intermediate pixel from previous input field) + t × (intermediate pixel from subsequent input field)

FIG. 8 is a diagram showing the outline of the operation of the interpolator 140 when the motion flag is not set. In this case, it is determined that the pixels currently being output do not form a moving part of the image, and therefore all vertical frequency bands are available (as shown in a and b of FIG. 5). Therefore, a full-bandwidth interpolation filter 630 will be applied to the interleaved pixels of the two input fields 320 and 330 to produce output pixels.

Although it is conceivable to use temporal weighting in the filter configuration shown in FIG. 8, it is not actually adopted in this embodiment for several reasons. (1) As the weightings of the two sets of pixels change, the filter must be spatially deformed to maintain a unity gain overall. Although such a modification is possible, a new constraint is added to the design of the filter, and the performance (for example, bandwidth) of the filter deteriorates. (2) With temporal weighting, as the temporal position of the output field approaches one of the two input fields, half of the pixel values dominate the filter. As a result, the bandwidth of the filter is halved. (3) Furthermore, since the time position of the output field with respect to the input field changes, the bandwidth decreases periodically as described in (2) above. This causes a periodic change in the apparent clarity of the murky object with a period of 10 Hz (in the case of conversion from 60 Hz to 50 Hz).

FIG. 9 is a schematic block diagram of the interpolator 140. The set of video fields consists of two interpolators 700, 7
10 and the current output field is interpolated from the preceding and succeeding input fields, respectively. Each interpolator 70
0 and 710 selectively operate according to the filter configurations of FIGS. 7 and 8.

The motion flag for each output pixel is provided to a "static" coefficient memory 720 and through an inversion gate 730 to a "dynamic" coefficient memory 740. An appropriate set of filter coefficients is output from one of the coefficient memories 720, 740 and used to interpolate output pixels depending on the state of the motion flag. The selected coefficient set is supplied to the interpolators 700 and 710 via the parallel-serial converter 750.

Each interpolator 700, 710 combines a number of pixels from each input field according to the selected set of filter coefficients. The output of the interpolator is multiplied by the weighting factors of the two multipliers 760 and 770, and the output from this is added by the adder 800. A weighting factor of either 0.5 or t (temporal position of the output field) is controlled by the parallel-serial converter 7 according to the control of the motion flag.
It is supplied from 80 to the multiplier 760. Similarly, a weighting factor of either 0.5 or (1-t) is again provided to the multiplier 770 by the parallel-to-serial converter 790 under the control of the motion flag.

When using t and (1-t) as the weighting factors, the interpolator 140 operates according to the filtering scheme shown in FIG. Are generated and combined using temporal weighting. When using the "hysterical" coefficients with their respective 0.5 weights, it is equivalent to the interleaved full bandwidth filter configuration shown in FIG. Interpolator 140 thus selectively operates as one of two different interpolators depending on the motion flag.

Various refinements of the above-described embodiments may alleviate the potential problem of a suddenly moving stationary object being brought into focus when the full bandwidth filter 630 is switched to the active state. . For example: (1) Between two output fields, one formed using a "static" (full bandwidth) interpolator and the other formed using a "dynamic" (half bandwidth) interpolator Cross-fade (connecting signals smoothly at edit points)
It can be performed. (2) Before using the full bandwidth filter, one or two intermediate bandwidth (between full bandwidth and half bandwidth) filters can be used after the motion has settled to zero.

[0062]

According to the present invention, there is provided means for detecting whether each pixel of the output image represents a dynamic part or a static part of the output image, and the information obtained through this means is detected. , For example, to adapt the interpolation method used to generate this pixel. Therefore, if the input image consists of a series of fields of interlaced video signals, information loss can be well compensated when the image moves vertically by an odd number of video lines between successive fields.

[Brief description of drawings]

FIG. 1 is a block diagram showing an outline of a motion compensation television system conversion device.

FIG. 2 is a block diagram showing an outline of operation of a motion vector selector.

FIG. 3 is a block diagram showing an outline of a motion vector selector.

FIG. 4 is a diagram showing an outline of a progressive scanning video frame.

FIG. 5 shows Example 1 of two consecutive video fields in an interlaced relationship.

FIG. 6 is a diagram showing an example 2 of two continuous video fields having an interlaced relationship.

FIG. 7 is a diagram showing an outline of operation of a motion compensation interpolator when a motion flag is set.

FIG. 8 is a diagram showing an outline of the operation of the motion compensation interpolator when the motion flag is not set.

FIG. 9 is a block diagram showing an outline of a motion compensation interpolator.

[Description of Reference Signs] 140 motion compensation interpolator 210 vector estimator 220 vector reducer 230 motion vector selector (vector selection means) 240 vector post-processor (motion flag generation means) 470 block matching comparator (interblock correlation detection) Means) 500 Comparator (motion detection means) 780,790 Parallel-serial converter (pixel selection means)

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Simon Matthew Mans United Kingdom Sally, Guildford Denzil Road 2 (72) Inventor Martin Rex Dorricot United Kingdom Hampshire, Basing Stoke, Basing, Ringfield Claus 6 (56) Reference Document JP-A-3-62690 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N ^7/ 00-7/68 H04N 11/00-11/22

Claims (5)

(57) [Claims] 1. A motion compensation video signal processing apparatus for generating a motion vector representing an image motion between a pair of input images, and forming an output image by interpolation for compensating the motion. For each input image pair,
Performs half-bandwidth interpolation to generate two intermediate pixel values
And then weight these two intermediate pixel values
And combine them to produce pixels in the output field.
Steps and interleave the input image pair when there is no image motion
And then all the interleaved pixels.
By performing the bandwidth interpolation process of the output field
Movement, characterized in that it comprises a means for generating a pixel
Compensation video signal processor . Wherein said input image pair, be composed of two successive fields of an interlaced input video signal
The motion-compensated video signal processing device according to claim 1. Wherein said output image is a motion compensated video signal processing apparatus <br/> according to claim 1, characterized in that it consists of fields of interlaced output video signal. 4. A motion-compensated video signal processing device for generating a motion vector representing a motion of an image between a pair of input images and forming an output image by interpolation for compensating the motion, comprising a plurality of motion-compensated pixel interpolators. select one from the plurality of motion compensation pixel interpolator in accordance with the motion detection of the image between the input image pair comprises a means for performing pixel interpolation of an output image from the input image pair
The motion-compensated video signal processor according to claim 1, characterized in that
Processing equipment. 5. An output image is generated by motion compensation interpolation.
Motion vector, which represents the motion of the image between a pair of input images.
A video signal processing method for generating a tor, when there is an image motion, for each of the input image pairs,
Performs half-bandwidth interpolation to generate two intermediate pixel values
And then weight these two intermediate pixel values
However, they can be combined to produce pixels in the output field.
And when there is no image motion, the input image pair is interleaved.
And, all Thereafter, the interleaved pixels
By performing the bandwidth interpolation process of the output field
Patent in that it comprises generating a pixel, the steps of
Video signal processing method .