CN101540902B - Method and device for scaling motion vectors, and method and system for coding/decoding - Google Patents
Method and device for scaling motion vectors, and method and system for coding/decoding Download PDFInfo
- Publication number
- CN101540902B CN101540902B CN 200810102353 CN200810102353A CN101540902B CN 101540902 B CN101540902 B CN 101540902B CN 200810102353 CN200810102353 CN 200810102353 CN 200810102353 A CN200810102353 A CN 200810102353A CN 101540902 B CN101540902 B CN 101540902B
- Authority
- CN
- China
- Prior art keywords
- motion vector
- carried out
- coordinate
- motion
- offset amount
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
- H04N19/112—Selection of coding mode or of prediction mode according to a given display mode, e.g. for interlaced or progressive display mode
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/513—Processing of motion vectors
- H04N19/517—Processing of motion vectors by encoding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/523—Motion estimation or motion compensation with sub-pixel accuracy
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
- H04N19/53—Multi-resolution motion estimation; Hierarchical motion estimation
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The embodiment of the invention relates to a method and a device for scaling motion vectors, and a method and a system for coding/decoding. The method for scaling motion vector comprises that: according to the amount of coordinate offset between a bottom field and a top field of an interlaced scanned image, coordinate transformation is performed on a first motion vector so as to obtain a second motion vector; the second motion vector is scaled so as to obtain a third motion vector; and according to the amount of coordinate offset, coordinate inverse transformation is performed to the third motion vector so as to obtain a fourth motion vector. The embodiment of the invention performs half-pixel compensation or coordinate transformation on the motion vectors before and after being scaled according to the offset of a top-field image and a bottom-field image in spatial positions, transforms a coordinate into a coordinate system reflecting the actual position relation of the images, and then scales the motion vector, thereby having the advantages of according with motion models better, guaranteeing the accuracy of scaling the motion vectors, raising the precision of predicting the motion vectors and improving the efficiency of coding and compressing field images.
Description
Technical field
The present invention relates to multimedia technology field, especially a kind of Zoom method of motion vector and device, decoding method and system.
Background technology
In the video coding compression standard, adopted the motion-vector prediction technology to reduce the required overhead bit of transmitting moving vector.The motion-vector prediction technology comprises spatial domain motion-vector prediction and time domain motion-vector prediction.In the motion-vector prediction of spatial domain, can be according to predicting the motion vector of current macro with current macro motion vector of adjacent macroblocks on the locus; And in the time domain motion-vector prediction, then can be according to predicting the motion vector of current macro with the motion vector of current macro adjacent or close macro block on the time shaft position.
Still be the time domain motion-vector prediction for the spatial domain motion-vector prediction no matter, the motion vector convergent-divergent is generally to adopt in the motion-vector prediction technology.Motion vector is one of most important data in the compressed bit stream, and the accuracy of motion vector convergent-divergent influences the precision of motion vector, and further influences the video coding compression performance.In general, the motion vector convergent-divergent is followed rigid body straight line uniform motion model, carries out corresponding convergent-divergent according to the time domain distance and handles.
In general, video sequence comprises sequence and interlaced line by line.The image of forming video sequence comprises that the two field picture of lining by line scan is that progressive scanning picture and interleaved field picture are horizontally interlaced image, 1 frame can be divided into 2, promptly there are a time difference in field, top and field, the end between field, top and the end field picture, and this time difference equals half of frame period.The field, top is made up of the odd-numbered line pixel, and field, the end is made up of the even number line pixel.Sequence only is made up of two field picture line by line, and interlacedly can comprise two field picture and field picture.
In the prior art, for two field picture and field picture, all carry out identical convergent-divergent and handle.But, because field picture is different from two field picture, there is the skew (being offset pixel half) on the locus between the field, top of field picture and the field, the end, adopt the motion vector convergent-divergent identical to handle the accuracy that can reduce the motion vector convergent-divergent, caused reduction the encoding compression efficient of field picture with the two field picture convergent-divergent.
Summary of the invention
The embodiment of the invention provides a kind of Zoom method and device, decoding method and system of motion vector, in order to improve the accuracy of motion vector convergent-divergent, improves the encoding compression efficient of field picture.
The embodiment of the invention provides a kind of Zoom method of motion vector, comprising:
According to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, obtain second motion vector;
Described second motion vector is carried out convergent-divergent, obtain the 3rd motion vector;
According to described coordinate offset amount described the 3rd motion vector is carried out the coordinate inverse transformation, obtain the 4th motion vector.
The embodiment of the invention provides a kind of device for zooming of motion vector, comprising:
First conversion module is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector;
Zoom module is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector;
Second conversion module is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
The embodiment of the invention provides a kind of coding method, comprising:
In cataloged procedure, when carrying out the motion-vector prediction processing, the step of motion vector being carried out convergent-divergent comprises:
According to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, obtain second motion vector;
Described second motion vector is carried out convergent-divergent, obtain the 3rd motion vector;
According to described coordinate offset amount described the 3rd motion vector is carried out the coordinate inverse transformation, obtain the 4th motion vector;
Described the 4th motion vector is used for motion-vector prediction to be handled.
The embodiment of the invention provides a kind of coding/decoding method, comprising:
In decode procedure, when carrying out the motion-vector prediction processing, the step of motion vector being carried out convergent-divergent comprises:
According to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, obtain second motion vector;
Described second motion vector is carried out convergent-divergent, obtain the 3rd motion vector;
According to described coordinate offset amount described the 3rd motion vector is carried out the coordinate inverse transformation, obtain the 4th motion vector;
Described the 4th motion vector is used for motion-vector prediction to be handled.
The embodiment of the invention provides a kind of coded system, comprising:
First device for zooming is used for first motion vector is carried out convergent-divergent, obtains the 4th motion vector;
First prediction unit is used for carrying out motion-vector prediction at the 4th motion vector that cataloged procedure obtains according to described first device for zooming and handles;
Wherein, described first device for zooming comprises:
First conversion module is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector;
Zoom module is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector;
Second conversion module is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
The embodiment of the invention provides a kind of decode system, comprising:
Second device for zooming is used for first motion vector is carried out convergent-divergent, obtains the 4th motion vector;
Second prediction unit is used for carrying out motion-vector prediction at the 4th motion vector that decode procedure obtains according to described second device for zooming and handles;
Wherein, described second device for zooming comprises:
First conversion module is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector;
Zoom module is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector;
Second conversion module is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
As shown from the above technical solution, embodiment of the invention skew on the locus according to top field picture and end field picture, motion vector before and after the convergent-divergent is carried out the half-pix compensation, carry out coordinate transform in other words, transform under the coordinate system of actual position relation between the reflection field picture, carry out the convergent-divergent of motion vector and handle, meet motion model more, guarantee the accuracy of motion vector convergent-divergent, improved the motion-vector prediction precision, thereby improved the encoding compression efficient of field picture.
Below in conjunction with accompanying drawing specific embodiments of the invention are described in further detail.
Description of drawings
Fig. 1 is the schematic flow sheet of first embodiment of the Zoom method of motion vector of the present invention;
Fig. 2 is the convergent-divergent schematic diagram of motion vector among first embodiment of the Zoom method of motion vector of the present invention;
Fig. 3 is the schematic flow sheet of second embodiment of the Zoom method of motion vector of the present invention;
Fig. 4 is the convergent-divergent schematic diagram of motion vector among second embodiment of the Zoom method of motion vector of the present invention;
Fig. 5 is the structural representation of the device for zooming embodiment of motion vector of the present invention;
Fig. 6 is the schematic flow sheet of coding method embodiment of the present invention;
Fig. 7 is the schematic flow sheet of coding/decoding method embodiment of the present invention;
Fig. 8 is the structural representation of coded system embodiment of the present invention;
Fig. 9 is the structural representation of decode system embodiment of the present invention.
Embodiment
The Zoom method of the motion vector of the embodiment of the invention carries out coordinate transform according to the coordinate offset amount between end field picture and the top field picture to first motion vector, can obtain second motion vector; Second motion vector is carried out convergent-divergent handle, thereby can obtain the 3rd motion vector; According to above-mentioned coordinate offset amount the 3rd motion vector is carried out the coordinate inverse transformation, just can obtain the 4th motion vector, for carrying out follow-up motion-vector prediction.The Zoom method of the motion vector of the embodiment of the invention can guarantee the accuracy of motion vector convergent-divergent, thereby improves the motion-vector prediction precision, and the encoding compression efficient of field picture.
As shown in Figure 1, be the schematic flow sheet of first embodiment of the Zoom method of motion vector of the present invention.Present embodiment is applicable to the motion-vector prediction that carries out based on the spatial domain, may further comprise the steps:
Suppose that scaled motion vector is MV1, the motion vector behind the convergent-divergent is MV2.
Between step 101, top field picture and the end field picture coordinate offset amount delt.
For horizontally interlaced image, there are the skew of half-pixel in vertical direction in top field picture and end field picture, corresponding coordinate offset amount delt=0.5, and unit is whole pixel (pixel);
Scaled MV1 is the motion vector under the former coordinate system; MV1 ' after the corresponding coordinate transform is the motion vector under the new coordinate system.The operation of coordinate transform specifically comprises following situation (only the vertical component of MV1 being carried out coordinate transform):
When motion vector MV1 by top-＞bottom (top point to a field, the end), MV1 '=MV1+delt;
When motion vector MV1 by bottom-＞top (field, the end point to top), MV1 '=MV1-delt;
When motion vector MV1 by top-＞top (top point to top), MV1 '=MV1;
When motion vector MV1 by bottom-＞bottom (field, the end point to field, the end), MV1 '=MV1.
From above-mentioned situation as can be seen, when the parity of the field picture at the starting point of motion vector and terminal point place not simultaneously, need to consider the influence of coordinate offset delt;
Wherein, MV2 '=scale * MV1 '; In the formula, scale is a zoom factor, and it is relevant with the time gap of the motion vector of convergent-divergent front and back, and this time gap is that motion vector is projected in the length on the time shaft.
Motion vector MV2 ' behind the convergent-divergent is the motion vector under the new coordinate system; MV2 after the corresponding coordinate inverse transformation is the motion vector under the former coordinate system.The operation of coordinate transform specifically comprises following situation (only the vertical component of MV2 ' being carried out coordinate transform):
When motion vector MV2 by top-＞bottom (top point to a field, the end), MV2=MV2 '-delt;
When motion vector MV2 by bottom-＞top (field, the end point to top), MV2=MV2 '+delt;
When motion vector MV2 by top-＞top (top point to top), MV2=MV2 ';
When motion vector MV2 by bottom-＞bottom (field, the end point to field, the end), MV2=MV2 '.
From above-mentioned situation as can be seen, when the parity of the field picture at the starting point of motion vector and terminal point place not simultaneously, need to consider the influence of coordinate offset delt.
Present embodiment is the offset d elt on the locus according to top field picture and end field picture, its unit is identical with the unit of motion vector, motion vector MV1 before and after the convergent-divergent and MV2 ' are carried out the half-pix compensation, carry out coordinate transform in other words, transform under the coordinate system O ' of actual position relation between the reflection field picture, carrying out the convergent-divergent of motion vector handles, meet motion model more, guaranteed the accuracy of motion vector convergent-divergent, improve the motion-vector prediction precision, thereby improved the encoding compression efficient of field picture.
As shown in Figure 2, be the convergent-divergent schematic diagram of motion vector among first embodiment of the Zoom method of motion vector of the present invention.Original coordinates is O, and new coordinate is O ', and filled circles is represented the top field picture, and open circles is represented end field picture, and both difference are the position difference of field, the end.Under new coordinate system, exist the half-pixel on the vertical direction poor between field, top and the field, the end.New coordinate system has reflected the actual position relation between the field picture, and the convergent-divergent of motion vector should carry out under new coordinate system.Among the figure, time express time axle; Y axle remarked pixel position in vertical direction; Top represents the top field picture; Bottom represents end field picture; Delt is the coordinate offset amount; MV1 and MV2 are respectively the motion vector under the coordinate system O; MV1 ' and MV2 ' are respectively MV1 and MV2 and transform to motion vector under the coordinate system O '; Scale is a zoom factor, under illustrated case, and scale=3.Under new coordinate system, motion vector MV2 ' behind the convergent-divergent and scaled motion vector MV1 ' are overlapping.
Further, in the present embodiment, in step 101, if field picture has been adopted interpolation processing, so corresponding coordinate offset amount is calculated as follows: if adopt 1/4 interpolation, then coordinate offset amount delt=0.5 * 4=2, unit is 1/4 pixel, and the expression half-pixel is equivalent to 2 1/4 samples.For the interpolation method of other precision, can analogize.
Further, in the present embodiment, when the motion vector behind the convergent-divergent and the motion vector direction before the convergent-divergent not simultaneously, zoom factor scale related in the step 103 can be negative.
In the present embodiment, what above-mentioned motion vector was described is the motion of pixel.Further, motion vector can also be used to description block (Block) or macro block (Macro Block, motion MB).Piece and macro block are made up of the pixel of some.Macro block is made up of 16 * 16 pixels, and macro block can be divided into piece.
As shown in Figure 3, be the schematic flow sheet of second embodiment of the Zoom method of motion vector of the present invention.Present embodiment is applicable to the motion-vector prediction that carries out based on time domain, compares with a last embodiment, and the step 103 among the last embodiment item becomes:
Step 103 ', to the MV1 ' after the coordinate transform, carry out linear scale according to the rigid body uniform rectilinear motion model along the direction parallel with MV1 ', obtain motion vector MV2 '.
The starting point of motion vector MV2 ' behind scaled motion vector MV1 ' and the convergent-divergent is positioned on the different time-domain positions, but on the position, spatial domain corresponding relation is arranged.
As shown in Figure 4, be the convergent-divergent schematic diagram of motion vector among second embodiment of the Zoom method of motion vector of the present invention.Original coordinates is O, and new coordinate is O ', and filled circles is represented the top field picture, and open circles is represented end field picture, and both difference are the position difference of field, the end.Under new coordinate system, exist the half-pixel on the vertical direction poor between field, top and the field, the end.New coordinate system has reflected the actual position relation between the field picture, and the convergent-divergent of motion vector should carry out under new coordinate system.Among the figure, time express time axle; Y axle remarked pixel position in vertical direction; Top represents the top field picture; Bottom represents end field picture; Delt is the coordinate offset amount; MV1 and MV2 are respectively the motion vector under the coordinate system O; MV1 ' and MV2 ' are respectively MV1 and MV2 and transform to motion vector under the coordinate system O '; Scale is a zoom factor, under illustrated case, and scale=1/3.Under new coordinate system, the motion vector MV2 ' behind the convergent-divergent is parallel with scaled motion vector MV1 '.
Among the Zoom method embodiment of the invention described above motion vector, the skew on the locus according to top field picture and end field picture, after motion vector before and after the convergent-divergent carried out the half-pix compensation, carrying out the convergent-divergent of motion vector again handles, can guarantee the accuracy of motion vector convergent-divergent, improve the motion-vector prediction precision, thereby improved the encoding compression efficient of field picture.
As shown in Figure 5, be the structural representation of the device for zooming embodiment of motion vector of the present invention.Present embodiment comprises first conversion module 10, Zoom module 20 and second conversion module 30 that connects in turn.Wherein, the coordinate offset amount that first conversion module 10 is used between end field picture and the top field picture is carried out coordinate transform to first motion vector, obtains second motion vector; Zoom module 20 is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector; Second conversion module 30 is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
Among the device for zooming embodiment of the invention described above motion vector, the skew on the locus of first conversion module and second conversion module according to top field picture and end field picture, after respectively the motion vector before and after the convergent-divergent being carried out the half-pix compensation, carry out the convergent-divergent processing of motion vector again by Zoom module, can guarantee the accuracy of motion vector convergent-divergent, improve the motion-vector prediction precision, thereby improved the encoding compression efficient of field picture.
Further, still as shown in Figure 5, first conversion module 10 can comprise interconnective first recognition unit 11 and first converter unit 12 in the present embodiment.Wherein, first recognition unit 11 is used to discern the direction of described first motion vector; First converter unit 12 is used for according to the recognition result of described first recognition unit 11 and described coordinate offset amount the vertical component of described first motion vector being carried out coordinate transform, obtains second motion vector.Second conversion module 30 can comprise interconnective second recognition unit 31 and second converter unit 32.Second recognition unit 31 is used to discern the direction of described the 3rd motion vector; Second converter unit 32 is used for according to the recognition result of described second recognition unit 31 and described coordinate offset amount the vertical component of described the 3rd motion vector being carried out coordinate transform, obtains the 4th motion vector.
As shown in Figure 6, be the schematic flow sheet of coding method embodiment of the present invention.Present embodiment is in cataloged procedure, and when carrying out the motion-vector prediction processing, the step of motion vector being carried out convergent-divergent comprises:
Present embodiment when carrying out the motion-vector prediction processing, carries out convergent-divergent to motion vector and handles in cataloged procedure, promptly according to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, can obtain second motion vector; Second motion vector is carried out convergent-divergent handle, thereby can obtain the 3rd motion vector; According to above-mentioned coordinate offset amount the 3rd motion vector is carried out the coordinate inverse transformation, just can obtain the 4th motion vector; The 4th motion vector is used for motion-vector prediction to be handled.Present embodiment has guaranteed the accuracy of motion vector convergent-divergent, thereby has improved the motion-vector prediction precision, and the encoding compression efficient of field picture, has further improved the efficient of coding.
As shown in Figure 7, be the schematic flow sheet of coding/decoding method embodiment of the present invention.Present embodiment is in decode procedure, and when carrying out the motion-vector prediction processing, the step of motion vector being carried out convergent-divergent comprises:
Present embodiment when carrying out the motion-vector prediction processing, carries out convergent-divergent to motion vector and handles in decode procedure, promptly according to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, can obtain second motion vector; Second motion vector is carried out convergent-divergent handle, thereby can obtain the 3rd motion vector; According to above-mentioned coordinate offset amount the 3rd motion vector is carried out the coordinate inverse transformation, just can obtain the 4th motion vector; The 4th motion vector is used for motion-vector prediction to be handled.Present embodiment has guaranteed the accuracy of motion vector convergent-divergent, thereby has improved the motion-vector prediction precision, and the encoding compression efficient of field picture, has further improved the efficient of decoding.
As shown in Figure 8, be the structural representation of coded system embodiment of the present invention.Present embodiment comprises interconnective first device for zooming 1 and first prediction unit 2.Wherein, first device for zooming 1 is used for first motion vector is carried out convergent-divergent, obtains the 4th motion vector; First prediction unit 2 is used for carrying out motion-vector prediction at the 4th motion vector that cataloged procedure obtains according to first device for zooming 1 to be handled.Wherein, first device for zooming 1 can comprise first conversion module 10, Zoom module 20 and second conversion module 30 that connects in turn.Wherein, first conversion module 10 is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector; Zoom module 20 is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector; Second conversion module 30 is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
Among the invention described above coded system embodiment, the skew on the locus of first conversion module in first device for zooming and second conversion module according to top field picture and end field picture, after respectively the motion vector before and after the convergent-divergent being carried out the half-pix compensation, carry out the convergent-divergent processing of motion vector again by Zoom module, handle according to the motion-vector prediction that the 4th motion vector carries out in the cataloged procedure by first prediction unit at last, can guarantee the accuracy of motion vector convergent-divergent, improved the motion-vector prediction precision, thereby improved the encoding compression efficient of field picture, further improved the efficient of coding.
As shown in Figure 9, be the structural representation of decode system embodiment of the present invention.Present embodiment comprises interconnective second device for zooming 3 and second prediction unit 4.Wherein, first device for zooming 3 is used for first motion vector is carried out convergent-divergent, obtains the 4th motion vector; First prediction unit 4 is used for carrying out motion-vector prediction at the 4th motion vector that decode procedure obtains according to second device for zooming 3 to be handled.Wherein, first device for zooming 3 can comprise first conversion module 10, Zoom module 20 and second conversion module 30 that connects in turn.Wherein, first conversion module 10 is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector; Zoom module 20 is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector; Second conversion module 30 is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
Among the invention described above decode system embodiment, the skew on the locus of first conversion module in second device for zooming and second conversion module according to top field picture and end field picture, after respectively the motion vector before and after the convergent-divergent being carried out the half-pix compensation, carry out the convergent-divergent processing of motion vector again by Zoom module, handle according to the motion-vector prediction that the 4th motion vector carries out in the decode procedure by second prediction unit at last, can guarantee the accuracy of motion vector convergent-divergent, improved the motion-vector prediction precision, thereby improved the encoding compression efficient of field picture, further improved the efficient of decoding.
In the above embodiment of the present invention, there are the fact of the position deviation on the vertical direction in top field picture and end field picture, have influenced code efficiency.In the video coding framework, to consider that in motion vector prediction process other also will consider this point as interpolation, loop filtering etc. the above-mentioned this point.By improving interpolation, loop filtering, improve precision, thereby improve encoding compression efficient.
Through the above description of the embodiments, those of ordinary skill in the art can be well understood to the present invention and can realize by hardware, also can realize by the mode that software adds essential general hardware platform.Based on such understanding, technical scheme of the present invention can be written as computer executable program, and computer equipment can be realized method provided by the invention by carrying out this program.This computer executable program can be stored in the non-volatile memory medium and (comprise CD-ROM, USB flash disk, portable hard drive etc.).
It should be noted that at last: above embodiment only in order to technical scheme of the present invention to be described, is not intended to limit; Although with reference to previous embodiment the present invention is had been described in detail, those of ordinary skill in the art is to be understood that: it still can be made amendment to the technical scheme that aforementioned each embodiment put down in writing, and perhaps part technical characterictic wherein is equal to replacement; And these modifications or replacement do not make the essence of appropriate technical solution break away from the spirit and scope of various embodiments of the present invention technical scheme.
Claims (18)
1. the Zoom method of a motion vector is characterized in that comprising:
According to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, obtain second motion vector;
Described second motion vector is carried out convergent-divergent, obtain the 3rd motion vector;
According to described coordinate offset amount described the 3rd motion vector is carried out the coordinate inverse transformation, obtain the 4th motion vector.
2. the Zoom method of motion vector according to claim 1 is characterized in that the unit of described coordinate offset amount is identical with the unit of motion vector.
3. the Zoom method of motion vector according to claim 2 is characterized in that described coordinate offset amount is 0.5 if field picture is not made interpolation processing, and unit is a pixel.
4. the Zoom method of motion vector according to claim 2 is characterized in that described coordinate offset amount equals 0.5 * 4 and also promptly equals 2 if field picture is made 1/4 interpolation processing, and unit is 1/4 pixel.
5. according to the Zoom method of claim 1,2,3 or 4 described motion vectors, it is characterized in that describedly first motion vector is carried out coordinate transform specifically comprising: the vertical component to first motion vector is carried out coordinate transform.
6. the Zoom method of motion vector according to claim 5 is characterized in that described vertical component to first motion vector carries out coordinate transform and specifically comprise:
If described first motion vector is by a sensing field, the end, top, then the vertical component with described first motion vector adds the above coordinate offset amount;
If described first motion vector points to the field, top by field, the end, then the vertical component with described first motion vector deducts described coordinate offset amount;
Point to field, the end as if described first motion vector by a field, sensing top, top or by field, the end, then the vertical component of described first motion vector is not done conversion.
7. according to the Zoom method of claim 1,2,3 or 4 described motion vectors, it is characterized in that describedly described second motion vector is carried out convergent-divergent specifically comprising: described second motion vector is carried out linear scale according to the rigid body uniform rectilinear motion model.
8. the Zoom method of motion vector according to claim 7 is characterized in that describedly according to the rigid body uniform rectilinear motion model described second motion vector being carried out linear scale and specifically comprising: along the direction of described second motion vector or the direction parallel with described second motion vector described second motion vector is carried out linear scale according to the rigid body uniform rectilinear motion model.
9. the Zoom method of motion vector according to claim 8, the zoom factor that it is characterized in that linear scale depends on that described first motion vector and described the 4th motion vector are projected in the length on the time shaft, if when the direction of described first motion vector and described the 4th motion vector was inequality, described zoom factor was a negative.
10. according to the Zoom method of claim 1,2,3 or 4 described motion vectors, it is characterized in that describedly described the 3rd motion vector is carried out the coordinate inverse transformation specifically comprising: the vertical component to described the 3rd motion vector is carried out the coordinate inverse transformation.
11. the Zoom method of motion vector according to claim 10 is characterized in that described vertical component to described the 3rd motion vector carries out the coordinate inverse transformation and specifically comprise:
If described the 3rd motion vector is by a sensing field, the end, top, then the vertical component with described the 3rd motion vector deducts described coordinate offset amount;
If described the 3rd motion vector points to the field, top by field, the end, then the vertical component with described the 3rd motion vector adds the above coordinate offset amount;
Point to field, the end as if described the 3rd motion vector by a field, sensing top, top or by field, the end, then the vertical component of described the 3rd motion vector is not done conversion.
12. the device for zooming of a motion vector is characterized in that comprising:
First conversion module is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector;
Zoom module is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector;
Second conversion module is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
13. the device for zooming of motion vector according to claim 12 is characterized in that described first conversion module comprises:
First recognition unit is used to discern the direction of described first motion vector;
First converter unit is used for according to the recognition result and the described coordinate offset amount of described first recognition unit vertical component of described first motion vector being carried out coordinate transform, obtains second motion vector.
14. the device for zooming of motion vector according to claim 12 is characterized in that described second conversion module comprises:
Second recognition unit is used to discern the direction of described the 3rd motion vector;
Second converter unit is used for according to the recognition result and the described coordinate offset amount of described second recognition unit vertical component of described the 3rd motion vector being carried out coordinate transform, obtains the 4th motion vector.
15. a coding method is characterized in that, comprising:
In cataloged procedure, when carrying out the motion-vector prediction processing, the step of motion vector being carried out convergent-divergent comprises:
According to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, obtain second motion vector;
Described second motion vector is carried out convergent-divergent, obtain the 3rd motion vector;
According to described coordinate offset amount described the 3rd motion vector is carried out the coordinate inverse transformation, obtain the 4th motion vector;
Described the 4th motion vector is used for motion-vector prediction to be handled.
16. a coding/decoding method is characterized in that, comprising:
In decode procedure, when carrying out the motion-vector prediction processing, the step of motion vector being carried out convergent-divergent comprises:
According to the coordinate offset amount between end field picture and the top field picture first motion vector is carried out coordinate transform, obtain second motion vector;
Described second motion vector is carried out convergent-divergent, obtain the 3rd motion vector;
According to described coordinate offset amount described the 3rd motion vector is carried out the coordinate inverse transformation, obtain the 4th motion vector;
Described the 4th motion vector is used for motion-vector prediction to be handled.
17. a coded system is characterized in that, comprising:
First device for zooming is used for first motion vector is carried out convergent-divergent, obtains the 4th motion vector;
First prediction unit is used for carrying out motion-vector prediction at the 4th motion vector that cataloged procedure obtains according to described first device for zooming and handles;
Wherein, described first device for zooming comprises:
First conversion module is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector;
Zoom module is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector;
Second conversion module is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
18. a decode system is characterized in that, comprising:
Second device for zooming is used for first motion vector is carried out convergent-divergent, obtains the 4th motion vector;
Second prediction unit is used for carrying out motion-vector prediction at the 4th motion vector that decode procedure obtains according to described second device for zooming and handles;
Wherein, described second device for zooming comprises:
First conversion module is used for according to the coordinate offset amount between end field picture and the top field picture first motion vector being carried out coordinate transform, obtains second motion vector;
Zoom module is used for described second motion vector is carried out convergent-divergent, obtains the 3rd motion vector;
Second conversion module is used for according to described coordinate offset amount described the 3rd motion vector being carried out the coordinate inverse transformation, obtains the 4th motion vector.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810102353 CN101540902B (en) | 2008-03-20 | 2008-03-20 | Method and device for scaling motion vectors, and method and system for coding/decoding |
| PCT/CN2008/073864 WO2009114986A1 (en) | 2008-03-20 | 2008-12-30 | Method, device for scaling motion vector and encoding, decoding method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810102353 CN101540902B (en) | 2008-03-20 | 2008-03-20 | Method and device for scaling motion vectors, and method and system for coding/decoding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101540902A CN101540902A (en) | 2009-09-23 |
| CN101540902B true CN101540902B (en) | 2011-02-02 |
Family
ID=41090480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200810102353 Active CN101540902B (en) | 2008-03-20 | 2008-03-20 | Method and device for scaling motion vectors, and method and system for coding/decoding |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN101540902B (en) |
| WO (1) | WO2009114986A1 (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5786478B2 (en) * | 2011-06-15 | 2015-09-30 | 富士通株式会社 | Moving picture decoding apparatus, moving picture decoding method, and moving picture decoding program |
| RU2716563C2 (en) * | 2011-09-09 | 2020-03-12 | Кт Корпорейшен | Video decoding method |
| CN106331722B (en) | 2015-07-03 | 2019-04-26 | 华为技术有限公司 | Image prediction method and relevant device |
| CN107135397B (en) * | 2017-04-28 | 2018-07-06 | 中国科学技术大学 | A kind of panorama video code method and apparatus |
| CN109729352B (en) | 2017-10-27 | 2020-07-21 | 华为技术有限公司 | Method and device for determining motion vector of affine coding block |
| CN110868611B (en) * | 2018-08-27 | 2023-09-12 | 北京三星通信技术研究有限公司 | Video encoding and decoding methods and devices |
| CN109068140B (en) * | 2018-10-18 | 2021-06-22 | 北京奇艺世纪科技有限公司 | Method and device for determining motion vector in video coding and decoding equipment |
| CN112087626A (en) * | 2020-08-21 | 2020-12-15 | 西安万像电子科技有限公司 | Image processing method, device and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1142733A (en) * | 1995-08-08 | 1997-02-12 | 大宇电子株式会社 | Image processing system using feature point-based motion estimation |
| CN1189057A (en) * | 1996-10-31 | 1998-07-29 | 松下电器产业株式会社 | Motion image coding method and device and its coding programmed recording medium |
| US6430223B1 (en) * | 1997-11-01 | 2002-08-06 | Lg Electronics Inc. | Motion prediction apparatus and method |
| CN1520179A (en) * | 2003-02-03 | 2004-08-11 | 三星电子株式会社 | Method and appts. for coding/decoding interlaced scanning video signal |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4158442B2 (en) * | 2002-07-18 | 2008-10-01 | 富士通株式会社 | Video encoding method and video decoding method |
-
2008
- 2008-03-20 CN CN 200810102353 patent/CN101540902B/en active Active
- 2008-12-30 WO PCT/CN2008/073864 patent/WO2009114986A1/en active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1142733A (en) * | 1995-08-08 | 1997-02-12 | 大宇电子株式会社 | Image processing system using feature point-based motion estimation |
| CN1189057A (en) * | 1996-10-31 | 1998-07-29 | 松下电器产业株式会社 | Motion image coding method and device and its coding programmed recording medium |
| US6430223B1 (en) * | 1997-11-01 | 2002-08-06 | Lg Electronics Inc. | Motion prediction apparatus and method |
| CN1520179A (en) * | 2003-02-03 | 2004-08-11 | 三星电子株式会社 | Method and appts. for coding/decoding interlaced scanning video signal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101540902A (en) | 2009-09-23 |
| WO2009114986A1 (en) | 2009-09-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101540902B (en) | Method and device for scaling motion vectors, and method and system for coding/decoding | |
| US11683521B2 (en) | Representative motion information for temporal motion prediction in video encoding and decoding | |
| CN111093074A (en) | Motion candidate derivation based on multiple information in sub-block motion vector prediction | |
| CN110786013B (en) | Search region for motion vector refinement | |
| KR101683434B1 (en) | Image processing device, decoding method, intra-frame decoder, intra-frame decoding method, and intra-frame encoder | |
| WO2017204886A1 (en) | System and method for providing real-time super-resolution for compressed videos | |
| EP2405655B1 (en) | Boundary padding for unrestricted motion vectors | |
| CN102316317A (en) | Method and device for generating predicted value of picture | |
| US11805259B2 (en) | Non-affine blocks predicted from affine motion | |
| CN107360433B (en) | Inter-frame prediction coding method and device | |
| US20070223588A1 (en) | Method of Searching for a Global Motion Vector | |
| CN113709458B (en) | Displacement vector prediction method, device and equipment in video coding and decoding | |
| CN105491380A (en) | Method and apparatus for motion vector predictor derivation | |
| CN101257625B (en) | Method for indexing position in video decoder and video decoder | |
| US9420303B2 (en) | Method and apparatus for displacement vector component transformation in video coding and decoding | |
| US11330296B2 (en) | Systems and methods for encoding image data | |
| CN101194513B (en) | Up-down alternative motion vector | |
| CN1848960B (en) | Residual coding in compliance with a video standard using non-standardized vector quantization coder | |
| US20100220786A1 (en) | Method and apparatus for multiple reference picture motion estimation | |
| CN1283107C (en) | Quick movement prediction method and structure thereof | |
| TWI834727B (en) | Non-affine blocks predicted from affine motion | |
| JP2866325B2 (en) | Motion vector search device | |
| CN103888780B (en) | Moving picture encoding device, dynamic image decoding device, motion image encoding method and dynamic image decoding method | |
| JP2963369B2 (en) | Motion vector search device | |
| CN101720038A (en) | Motion compensating device and motion compensating method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |