CN103581504A - Motion compensated image processing device and image processing method - Google Patents
Motion compensated image processing device and image processing method Download PDFInfo
- Publication number
- CN103581504A CN103581504A CN201210251213.2A CN201210251213A CN103581504A CN 103581504 A CN103581504 A CN 103581504A CN 201210251213 A CN201210251213 A CN 201210251213A CN 103581504 A CN103581504 A CN 103581504A
- Authority
- CN
- China
- Prior art keywords
- motion
- vector
- image
- image block
- motion compensation
- 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.)
- Granted
Links
Images
Abstract
The invention relates to a motion compensated image processing device comprising an external memory, a cache memory, a motion compensation module, a judgment module and a retrieval module. The external memory is used for storing a reference picture associated with an image block. The motion compensation module is used for carrying out motion compensation for a pre-image block and the image block sequentially. The judgment module judges a motion vector of the image block relative to the reference picture when the motion compensation module carries out motion compensation on the pre-image block. The retrieval module retrieves a reference area, corresponding to the motion vector, in the reference picture to the cache memory from the external memory before the motion compensation module carries out motion compensation on the image block.
Description
Technical field
The present invention is relevant to image processing technique, and especially relevant to the technology of managing/using the memory in image processing system.
Background technology
Motion compensation (motion compensation) is a kind of mechanism that is widely used in dynamic video compression field.Animation expert body (moving picture experts group, MPEG) the block type motion compensation that standard was adopted of take is example, and picture to be encoded is split into the block that a plurality of sizes are 16 pixel * 16 pixels; For each block, encoder can be found out a reference zone the most similar from reference picture, and judges the motion-vector between a block and corresponding reference zone thereof.Except motion-vector, encoder also can judge the presentation content difference between a block and corresponding reference zone thereof, is called amount of redundancy (residual).After the factor of one block deduction motion-vector, more similar to its reference zone, amount of redundancy is less.In coding result, each block is to represent with its motion-vector and amount of redundancy.
Fig. 1 (A) is in order to represent the work schedule example of existing encoder when carrying out motion compensation.Encoder is the most similar reference zone of block A in the interior judgement of time section T 1 picture to be encoded first.Due in order to be normally stored in the external memory storage of encoder with the reference picture of picture to be encoded comparison, suppose to be distributed in block R1 and the block R2 in reference picture at the time section T 1 inner encoder judgement reference zone the most similar to block A in picture to be encoded.When the amount of redundancy producing corresponding to block A, encoder must capture from external memory storage the image data of block R1 and block R2.Time section T 2 represents that encoder sends the requirement of acquisition block R1 and wait for afterwards the time that external memory storage is responded.3 representatives of time section T are passed to block R1 from external memory storage the time of encoder.Time section T 4 represents that encoder sends the requirement of acquisition block R2 and wait for afterwards the time that external memory storage is responded.5 representatives of time section T are passed to block R2 from external memory storage the time of encoder.6 of time section T are to represent that encoder produces the time corresponding to the amount of redundancy of block A according to block R1 and block R2.The rest may be inferred, when encoder will produce amount of redundancy for each block in picture to be encoded, all must consume six periods such as T1 ~ T6.
Fig. 1 (B) is in order to represent the work schedule example of existing decoder when carrying out motion compensation.Similarly, in order to be normally stored in the external memory storage of decoder with the reference picture of picture comparison to be decoded.When wish is rebuild aforementioned block A, first decoder judges that according to the motion-vector in decoded result the reference zone that produces according to this this motion-vector is distributed in block R1 and the block R2 in reference picture in time section T 1 '.Therefore, decoder must be from the image data of its external memory storage acquisition block R1 and block R2.Time section T 2 ' represents that decoder sends the requirement of acquisition block R1 and wait for afterwards the time that external memory storage is responded.3 ' the representative of time section T is passed to block R1 from external memory storage the time of decoder.Time section T 4 ' represents that decoder sends the requirement of acquisition block R2 and wait for afterwards the time that external memory storage is responded.5 ' the representative of time section T is passed to block R2 from external memory storage the time of decoder.Time section T 6 ' is to represent that decoder is according to the time of block R1, block R2 and amount of redundancy reconstruction block A.The rest may be inferred, when decoder will be rebuild each block in picture to be decoded, all must consume six periods such as T1 ' ~ T6 '.
One of development trend of image system is constantly to increase picture dimension and resolution now, and therefore encoder is all required with higher operation efficiency.By above explanation, can be found out, the total length of time section T 1 ~ T6 and T1 ' ~ T6 ' is longer, and the operation efficiency of encoder is poorer.
Summary of the invention
For addressing the above problem, the present invention proposes a kind of motion compensation image processor and motion compensation image treatment method, when being imposed to motion compensation program, an image block simultaneously for next image block captures institute palpus reference data to memory cache from external memory storage in advance, reaches timesaving effect.In addition, according to image processor of the present invention and image treatment method, can adopt the concept that does not repeat to capture the data that had in memory cache, further promote by this treatment effeciency.
A specific embodiment according to the present invention is a kind of motion compensation image processor, wherein comprises external memory storage, memory cache, motion compensation module, judge module and acquisition module.External memory storage is in order to store the reference picture relevant to an image block.Motion compensation module is in order to sequentially to carry out motion compensation for a front image block and this image block.When this motion compensation module is carried out motion compensation to this front image block, judge module judges that this image block is with respect to the motion-vector of reference picture.Before this motion compensation module is carried out motion compensation to this image block, this acquisition module captures to memory cache from external memory storage with reference to the reference zone corresponding to this motion-vector in picture.
Another specific embodiment according to the present invention is the motion compensation image treatment method of a kind of cooperation one external memory storage and memory cache running.This external memory storage stores a reference picture relevant to an image block.First the method carries out a motion compensation step, and a front image block is carried out to motion compensation.Before this motion compensation step finishes, the method starts to carry out a determining step, judges that this image block is with respect to a motion-vector of this reference picture.Then, the method is carried out an acquisition step, and the reference zone corresponding to this motion-vector in this reference picture is captured to this memory cache from this external memory storage.Subsequently, the method is carried out another motion compensation step, and this image block is carried out to motion compensation.
Another specific embodiment according to the present invention is a kind of motion compensation image processor, wherein comprises external memory storage, memory cache, motion compensation module, judge module and acquisition module.External memory storage is in order to store a reference picture relevant to an image block.Motion compensation module is in order to carry out motion compensation for this image block.Judge module judges that this image block is with respect to a motion-vector of this reference picture.Acquisition module captures a remaining area that is not stored in this memory cache corresponding to a reference zone of this motion-vector to this memory cache from this external memory storage.
Another specific embodiment according to the present invention is the motion compensation image treatment method of a kind of cooperation one external memory storage and memory cache running.This external memory storage is in order to store a reference picture relevant to an image block.First the method carries out a determining step, judges that this image block is with respect to a motion-vector of this reference picture.Then, the method is carried out an acquisition step, from this external memory storage, a remaining area that is not stored in this memory cache corresponding to a reference zone of this motion-vector is captured to this memory cache from this external memory storage.Subsequently, the method is carried out a motion compensation step, according to this reference zone that is stored in this memory cache, for this image block, carries out motion compensation.
About the advantages and spirit of the present invention, can be further understood by following detailed description and accompanying drawings.
Accompanying drawing explanation
Fig. 1 (A) represents the work schedule example of existing encoder when carrying out motion compensation; Fig. 1 (B) represents the work schedule example of existing decoder when carrying out motion compensation.
Fig. 2 is according to the encoder block figure in one embodiment of the invention.
Fig. 3 is in order to represent according to the work schedule of encoder of the present invention.
Fig. 4 is the corresponding relation example of a reference picture and reference zone.
Fig. 5 (A) and Fig. 5 (B) are according to the decoder block figure in one embodiment of the invention.
Fig. 6 is according to the image processor calcspar in one embodiment of the invention.
Fig. 7 and Fig. 8 are according to the image treatment method flow chart in embodiments of the invention.
Main element symbol description
T1 ~ T9, T1 ' ~ T6 ': time section 200: encoder
300: decoder 400: image processor
21,31,41: external memory storage 22,32,42: memory cache
23,33,43: motion compensation module 24,34,44: judge module
25,35,45: acquisition module 36: amount of redundancy generation module
37: analysis module 37R1 ~ R6: the block in reference picture
X1, X2: reference zone S71 ~ S74, S81 ~ S86: process step
Embodiment
According to one embodiment of the invention, be the encoder 200 shown in Fig. 2, wherein comprise external memory storage 21, memory cache (cache) 22, motion compensation module 23, judge module 24 and acquisition module 25.In practical application, encoder 200 can be incorporated in various image processing systems or image playback equipment, also can independently exist.
As shown in Figure 2, because the bit element flow of picture to be encoded is provided respectively to motion compensation module 23 and judge module 24, make motion compensation module 23 and judge module 24 for different blocks, to carry out image processing simultaneously.In detail, suppose that motion compensation module 23 sequentially carries out motion compensation coded program for two image block B1 in picture to be encoded, B2 (not being shown in graphic), produce respectively amount of redundancy and motion-vector corresponding to image block B1, B2, as representing data after the coding of image block B1, B2.When 23 couples of image block B1 of motion compensation module carry out motion compensation coded program, judge module 24 can judge that image block B2 is with respect to the motion-vector of reference picture.Judge module 24 produces after the motion-vector of image block B2, acquisition module 25 is just according to this motion-vector, with reference to the reference zone corresponding to this motion-vector in picture, from external memory storage 21, capture to memory cache 22, while carrying out motion compensation coded program for 23 couples of image block B2 of motion compensation module.
Fig. 3 is in order to the work schedule of presentation code device 200.Time section T 1 represents that 23 couples of image block B1 of motion compensation module carry out the time of motion compensation coded program.Time section T 2 represents that judge module 24 is for generation of the time of the motion-vector of image block B2.Suppose that the motion-vector that judge module 24 produces points out to be distributed in block R1 and the block R2 in reference picture corresponding to the reference zone of image block B2.Time section T 3 represents that acquisition module 25 sends the time that the requirement of acquisition block R1 waits for that external memory storage 21 is responded afterwards.4 representatives of time section T are passed to block R1 from external memory storage 21 time of memory cache 22.Time section T 5 represents that acquisition module 25 sends the time that the requirement of acquisition block R2 waits for that external memory storage 21 is responded afterwards.6 representatives of time section T are passed to block R2 from external memory storage 21 time of memory cache 22.Time section T 7 and T8 represent that respectively motion compensation module 23 reads the time of block R1 and R2 from memory cache 22.9 of time section T represent that 23 couples of image block B2 of motion compensation module carry out the time of motion compensation coded program.
As shown in Figure 3, the time section T 2 in this embodiment, T3 and time section T 1 overlap completely, and time section T 4 also overlaps with time section T 1 part.Although each time section length in Fig. 3 is all only example, but compared to the conventional situation shown in Fig. 1 (A), because encoder 200 has shifted to an earlier date, start from the time of external memory storage 21 acquisition datas, in fact still can complete in the short period of time image block coded program.It should be noted that, it is identical with the initial time of time section T 1 that the initial time of time section T 2 can be advanced to, or even early than the initial time of time section T 1.
In practice, judge module 24 can, by directly comparing the motion-vector that image block B2 and this reference picture decide image block B2, can also produce a prediction motion-vector as the motion-vector of image block B2 according at least one adjacent block of image block B2 and/or at least one adjacent motion-vector.For example, if image block B1, B2 are adjacent one another are, judge module 24 can be inferred according to the motion-vector of the image block B1 having produced the motion-vector of image block B2.Compared to directly comparing image block and reference picture, by prediction, decide motion-vector conventionally comparatively to save time, can further promote the efficiency of encoder 200.Generally speaking, adjacent image block often has close motion-vector.Therefore, according to adjacent motion-vector as fundamentals of forecasting can obtain as a rule can be too not large with optimum value gap result.Should be noted that, the object of judge module 24 references is not limited to the motion-vector of image block B1, and the motion-vector of prediction image block B2 is also average after the weight by the motion-vector of adjacent a plurality of image blocks.In addition, judge module 24 also can for example, select to predict according to this data of its motion-vector according to the known properties of image block B2 (whether being the object boundary in image) elasticity.
From the above, the prediction motion-vector of image block B2 and the gap of a best motion-vector are conventionally not too large.Therefore, motion compensation module 23 can be designed to not consider the motion-vector predicted and the gap of best motion-vector, the motion-vector that namely no matter judge module 24 is predicted why, and motion compensation module 23 is all carried out motion compensation coded program according to this prediction motion-vector.Easy speech, motion compensation module 23 can adopt image block B2 and corresponding to the image difference between the reference zone of the motion-vector of this prediction the amount of redundancy as image block B2, and produce data after the coding that represents image block B2 according to this image difference and this motion-vector.
In another embodiment, before 23 couples of image block B2 of motion compensation module carry out motion compensation coded program, judge module 24 can be designed to further to judge that image block B2 is with respect to the actual motion-vector (namely aforementioned best motion-vector) of reference picture, and the motion-vector of comparison prediction and actual motion-vector.If the difference of prediction motion-vector and actual motion-vector is greater than a preset threshold, acquisition module 25 further in external memory storage 21 acquisition reference pictures corresponding to an actual reference zone of actual motion-vector, for motion compensation module 23 when the processing image block B2.Relatively, if the difference of prediction motion-vector and actual motion-vector is less than or equal to this preset threshold, acquisition module 25 does not just need again by external memory storage 21 acquisition datas, and motion compensation module 23 can be directly according to corresponding to prediction motion-vector and to have captured to the reference zone of memory cache 22 be that image block B2 produces data after coding.
Should be noted that, acquisition module 25 can be controlled this actual reference zone and directly from external memory storage 21, be captured to motion compensation module 23, not necessarily will pass through memory cache 22.In practical application, as long as judge module 24 has produced prediction motion-vector, acquisition module 25 just can start the reference zone corresponding to this prediction motion-vector to capture to memory cache 22, need not wait for that actual motion-vector produces.In addition,, because needs are again little from the probability of the actual reference zone of external memory storage 21 acquisition, generally speaking, even if must again capture above-mentioned actual reference zone once in a while, the code efficiency of encoder 200 is still better than prior art.
As mentioned above, adjacent image block often has close motion-vector.Therefore, adjacent image block is possibly corresponding to adjacent reference zone.Fig. 4 is the corresponding relation example of a reference picture and reference zone.In this example, the reference zone X1 of image block B1 is distributed in block R1, R2, R4, the R5 in reference picture, and the reference zone X2 of image block B2 is block R2, R3, R5, the R6 being distributed in reference picture.Just utilize a dynamic random access memory to realize the situation of external memory storage 21, be limited to the access features of dynamic random access memory, in order to allow motion compensation module 23 can process image block B1, acquisition module 25 probably must be by block R1, R2, R4, R5 complete acquisitions to memory cache 22 all, and can not only capture reference zone X1.
In an embodiment, before the motion-vector producing according to judge module 24 captures a reference zone, first acquisition module 25 judges in memory cache 22, whether there has been this all or part of reference zone.If there has been this whole reference zones in memory cache 22, just acquisition module 25 is no longer from any data of external memory storage 21 acquisition.If there has been this reference zone of part in memory cache 22,25 of acquisition modules are that a remaining area that is only not yet stored in memory cache 22 from external memory storage 21 with reference to region captures to memory cache 22, no longer repeat acquisition and have been stored in the data in memory cache 22.The situation that Fig. 4 was illustrated of take is example, in order to process image block B1, has had block R1, R2, R4, R5 in memory cache 22.If the motion-vector that judge module 24 produces points out that the reference zone X2 of image block B2 is distributed in block R2, R3, R5, the R6 in reference picture, acquisition module 25 will be only that image block B2 captures block R3, R6 from external memory storage 21, and not repeat to capture block R2, R5.Operating time and volume of transmitted data that this way that does not repeat acquisition data can further be saved encoder 200.
Be the decoder 300 shown in Fig. 5 (A) according to another embodiment of the present invention, wherein comprise external memory storage 31, memory cache 32, motion compensation module 33, judge module 34, acquisition module 35 and amount of redundancy generation module 36.In practical application, decoder 300 can be incorporated in various image processing systems or image playback equipment, also can independently exist.External memory storage 31 is in order to store with motion compensation module 33 processing or the relevant reference picture of the image block processed.More particularly, each block that this reference picture is picture to be decoded produces the comparison basis of motion-vector and amount of redundancy according to this at coding side.
As shown in Fig. 5 (A), the bit element flow of picture to be decoded is provided respectively to judge module 34 and amount of redundancy generation module 36.Amount of redundancy generation module 36 is in order to analyze this bit element flow, to obtain the image difference between an actual reference zone of an image block and the corresponding employing of coding side, the namely amount of redundancy of this image block.Suppose that motion compensation module 33 sequentially carries out motion compensation decoding program for two image block B1, B2 in picture to be decoded, reconstruct respectively image block B1, B2.When 33 couples of image block B1 of motion compensation module carry out motion compensation decoding program, judge module 34 is responsible for judgement image block B2 with respect to the motion-vector of reference picture.At judge module 34, produce after the motion-vector of image block B2, acquisition module 35 is just according to this motion-vector, with reference to the reference zone corresponding to this motion-vector in picture, from external memory storage 31, capture to memory cache 32, while carrying out motion compensation decoding program for 33 couples of image block B2 of motion compensation module.The image difference reconstructed image block B2 that motion compensation module 33 can draw according to this reference zone and amount of redundancy generation module 36.
Similarly, judge module 34 can produce a prediction motion-vector as the motion-vector of image block B2 according at least one adjacent block of image block B2 and/or at least one adjacent motion-vector.Motion compensation module 33 can be designed to not consider to predict the gap of motion-vector and actual motion-vector, the prediction motion-vector that namely no matter judge module 34 produces why, and motion compensation module 33 is all carried out motion compensation decoding program according to this prediction motion-vector.Compared to the conventional situation shown in Fig. 1 (B), because having shifted to an earlier date, decoder 300 starts from the time of external memory storage 31 acquisition datas, therefore can complete in the short period of time image block decoding program.
Fig. 5 (B) illustrates another embodiment of decoder 300.In this embodiment, the amount of redundancy generation module 36 in Fig. 5 (A) is replaced by an analysis module 37.Analysis module 37 is in order to analyze this bit element flow, to obtain an image block with respect to image difference and actual motion-vector between the actual reference zone of coding side employing.It should be noted that before analysis module 37 obtains this actual motion-vector, acquisition module 35 has started this reference zone to capture to memory cache 32.At analysis module 37, obtain after this actual motion-vector, judge module 34 is this prediction motion-vector and this actual motion-vector relatively.If the difference of prediction motion-vector and actual motion-vector is greater than a preset threshold, acquisition module 35 can be further from external memory storage 31 these actual reference zones of acquisition, for motion compensation module 33 when this image block of processing.Relatively, if the difference of prediction motion-vector and actual motion-vector is less than or equal to preset threshold, just having captured to the reference zone corresponding to predicting motion-vector of memory cache 32, the direct employing of motion compensation module 33 carries out motion compensation decoding program.
In practical application, previously the described concept that does not repeat acquisition also can be applicable in the decoder 300 shown in Fig. 5 (A) or Fig. 5 (B).
Be the image processor 400 shown in Fig. 6 according to another embodiment of the present invention, wherein comprise external memory storage 41, memory cache 42, motion compensation module 43, judge module 44 and acquisition module 45.In practical application, image processor 400 can be incorporated in various image processing systems or image playback equipment, also can independently exist.External memory storage 41 is in order to store a reference picture relevant to an image block.Motion compensation module 43 for example, in order to carry out a motion compensation program (a motion compensation coded program or a motion compensation decoding program) for this image block.Judge module 44 is in order to judge that this image block is with respect to a motion-vector of this reference picture.Acquisition module 45 is in order to confirm in this reference picture the reference zone corresponding to this motion-vector, and judges in memory cache 42, whether there has been this all or part of reference zone.
If there is not any part of this reference zone in memory cache 42, acquisition module 45 can capture to memory cache 42 from external memory storage 41 this reference zone is whole.If there has been this reference zone of part in memory cache 42, the remaining area that acquisition module 45 is only not yet stored in memory cache 42 from external memory storage 41 by this reference zone captures to memory cache 42, for motion compensation module 43 when processing this image block.Relatively, if there have been all parts of this reference zone in memory cache 42, acquisition module 45 can not capture again.By avoiding repetition acquisition data, image processor 400 can effectively be saved the operating time.
Be the image treatment method of a kind of cooperation one external memory storage and memory cache running according to another embodiment of the present invention, its flow chart as shown in Figure 7.This external memory storage is in order to store a reference picture relevant to an image block.First, step S71 is for to carry out a motion compensation program to a front image block.Step S72 is for to start before step S71 finishes to judge that this image block is with respect to a motion-vector of this reference picture.After step S72 completes, step S73 is performed, and the reference zone corresponding to this motion-vector in this reference picture is captured to this memory cache from this external memory storage.Step S74 carries out this motion compensation program to this image block.Several circuit operation flow change of previously having described when introducing phase encoder 200, decoder 300, also can be applied in the image treatment method that Fig. 7 illustrates, and its details repeats no more.
Be the image treatment method of a kind of cooperation one external memory storage and memory cache running according to another embodiment of the present invention, its flow chart as shown in Figure 8.This external memory storage is in order to store a reference picture relevant to an image block.First, step S81 is the motion-vector of this image block of judgement with respect to this reference picture.Step S82 is a reference zone of confirming in this reference picture corresponding to this motion-vector.Step S83 is for judging whether there has been this all or part of reference zone in this memory cache.If the judged result of step S83 points out not have in this memory cache any part of this reference zone, step S84 will be performed, with this external memory storage certainly by whole acquisitions of this reference zone to this memory cache.If the judged result of step S83 points out to have had in this memory cache this reference zone of part, step S85 will be performed, and with a remaining area that only from this external memory storage, this reference zone is not yet stored in to this memory cache, capture to this memory cache.If the judged result of step S83 points out to have had in this memory cache all parts of this reference zone, step S86 is directly performed, to carry out a motion compensation program according to this reference zone that is stored in this memory cache for this image block.Motion compensation program in step S86 can be a motion compensation coded program or a motion compensation decoding program.
As mentioned above, the above embodiment of the present invention proposes a kind of image processor and image treatment method, when being imposed to motion compensation program, an image block simultaneously for next image block captures institute palpus reference data to memory cache from external memory storage in advance, reaches timesaving effect.In addition, image processor according to the above embodiment of the present invention and image treatment method can adopt the concept that does not repeat to capture the data that had in memory cache, further promote by this treatment effeciency.
The above detailed description of preferred embodiments wishes more to know description feature of the present invention and spirit, and not with the above-mentioned preferred embodiment being disclosed, category of the present invention is limited.On the contrary, its objective is that hope can contain in the category of the scope of the claims of being arranged in of various changes and tool equality institute of the present invention wish application.
Claims (18)
1. a motion compensation image processor, comprises:
One external memory storage, in order to store a reference picture relevant to an image block;
One memory cache;
One motion compensation module, in order to sequentially to carry out motion compensation for a front image block and this image block;
One judge module judges that this image block is with respect to a motion-vector of this reference picture when this motion compensation module is carried out motion compensation to this front image block; And
One acquisition module, is used to this motion compensation module this image block is carried out before motion compensation, and the reference zone corresponding to this motion-vector in this reference picture is captured to this memory cache from this external memory storage.
2. image processor as claimed in claim 1, is characterized in that, this judge module is according to the type information of an adjacent block of adjacent this image block or this motion-vector of adjacent motion vector prediction.
3. image processor as claimed in claim 2, it is characterized in that, this motion compensation module judges the image difference between this image block and this reference zone, and produces according to this image difference and this motion-vector the rear data of a coding that represent this image block.
4. image processor as claimed in claim 2, is characterized in that, also comprises:
One amount of redundancy generation module, in order to analyze corresponding to data after a coding of this image block to obtain the image difference between this image block and an actual reference zone;
Wherein this motion compensation module is rebuild this image block according to this reference zone and this image difference.
5. image processor as claimed in claim 2, it is characterized in that, in this motion compensation module, this image block is carried out before motion compensation, this judge module also judges an actual motion-vector of this image block, and relatively this motion-vector and this actual motion-vector; Before this actual motion-vector produces, this acquisition module starts this reference zone acquisition to this memory cache; If the difference of this motion-vector and this actual motion-vector is greater than a preset threshold, this acquisition module is the actual reference zone corresponding to this actual motion-vector from this external memory storage acquisition.
6. image processor as claimed in claim 2, is characterized in that, further comprises:
One analysis module, in order to analyze corresponding to data after a coding of this image block to obtain this image block with respect to an image difference and an actual motion-vector of an actual reference zone;
Wherein, before this analysis module obtains this actual motion-vector, this acquisition module starts this reference zone acquisition to this memory cache; This judge module is relatively this prediction motion-vector and this actual motion-vector also, and when the difference of this prediction motion-vector and this actual motion-vector is greater than a preset threshold, this acquisition module captures this actual reference zone from this external memory storage.
7. image processor as claimed in claim 1, is characterized in that, the remaining area acquisition that this acquisition module is not stored in this memory cache from this external memory storage by this reference zone is to this fast memory.
8. a motion compensation image treatment method that coordinates an external memory storage and memory cache running, this external memory storage stores a reference picture relevant to an image block, and this image treatment method comprises:
(a) a front image block is carried out to motion compensation;
(b), before step (a) finishes, start to judge that this image block is with respect to a motion-vector of this reference picture;
(c) reference zone corresponding to this motion-vector in this reference picture is captured to this memory cache from this external memory storage; And
(d) this image block is carried out to motion compensation.
9. image treatment method as claimed in claim 8, is characterized in that, step (b) comprises according to the type information of an adjacent block of adjacent this image block or this motion-vector of adjacent motion vector prediction.
10. image treatment method as claimed in claim 9, is characterized in that, step (d) comprises:
Judge the image difference between this image block and this reference zone; And
According to this image difference and this motion-vector, produce the rear data of a coding that represent this image block.
11. image treatment methods as claimed in claim 9, is characterized in that, in step (d), also comprise before:
Analysis is encoded rear data to obtain the image difference between this image block and an actual reference zone corresponding to one of this image block;
Wherein step (d) comprises according to this reference zone and this image difference and rebuilds this image block.
12. image treatment methods as claimed in claim 9, is characterized in that, this image treatment method further comprises before in step (d):
Judge that this image block is with respect to an actual motion-vector of this reference picture;
Relatively this motion-vector and this actual motion-vector; And
If the difference of this motion-vector and this actual motion-vector is greater than a preset threshold, this external memory storage acquisition is corresponding to an actual reference zone of this actual motion-vector certainly;
Wherein, before this actual motion-vector produces, step (c) has started to carry out.
13. image treatment methods as claimed in claim 9, is characterized in that, in step (d), further comprise before:
Analysis is encoded rear data to obtain this image block with respect to an image difference and an actual motion-vector of an actual reference zone corresponding to one of this image block;
Relatively this prediction motion-vector and this actual motion-vector are to produce a comparative result; And
If this comparative result is greater than a preset threshold, this external memory storage acquisition is corresponding to an actual reference zone of this actual motion-vector certainly;
Wherein, before obtaining this actual motion-vector, step (c) has started to carry out.
14. image treatment methods as claimed in claim 8, is characterized in that, step (c) comprises
A remaining area acquisition that this reference zone is not stored in to this memory cache from this external memory storage is to this memory cache.
15. 1 kinds of motion compensation image processors, comprise:
One external memory storage, in order to store a reference picture relevant to an image block;
One memory cache;
One motion compensation module, in order to carry out motion compensation for this image block;
One judge module, judges that this image block is with respect to a motion-vector of this reference picture; And
One acquisition module, in order to capture a remaining area that is not stored in this memory cache corresponding to a reference zone of this motion-vector to this memory cache from this external memory storage.
16. image processors as claimed in claim 15, is characterized in that, this motion compensation module is carried out motion compensation coding or motion compensation decoding.
17. 1 kinds of motion compensation image treatment methods that coordinate an external memory storage and memory cache running, this external memory storage is in order to store a reference picture relevant to an image block, and this image treatment method comprises:
Judge that this image block is with respect to a motion-vector of this reference picture;
From this external memory storage, a remaining area that is not stored in this memory cache corresponding to a reference zone of this motion-vector is captured to this memory cache; And
According to this reference zone that is stored in this memory cache, for this image block, carry out motion compensation.
18. image treatment methods as claimed in claim 17, is characterized in that, motion compensation is motion compensation coding or motion compensation decoding.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210251213.2A CN103581504B (en) | 2012-07-20 | 2012-07-20 | Motion compensation image processor and image treatment method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210251213.2A CN103581504B (en) | 2012-07-20 | 2012-07-20 | Motion compensation image processor and image treatment method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103581504A true CN103581504A (en) | 2014-02-12 |
| CN103581504B CN103581504B (en) | 2017-06-20 |
Family
ID=50052353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210251213.2A Expired - Fee Related CN103581504B (en) | 2012-07-20 | 2012-07-20 | Motion compensation image processor and image treatment method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103581504B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10134A (en) * | 1853-10-18 | Cutting boots | ||
| CN101345026A (en) * | 2007-07-10 | 2009-01-14 | 联詠科技股份有限公司 | Frame data buffering apparatus and related frame data acquisition method |
| US7965773B1 (en) * | 2005-06-30 | 2011-06-21 | Advanced Micro Devices, Inc. | Macroblock cache |
-
2012
- 2012-07-20 CN CN201210251213.2A patent/CN103581504B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10134A (en) * | 1853-10-18 | Cutting boots | ||
| US7965773B1 (en) * | 2005-06-30 | 2011-06-21 | Advanced Micro Devices, Inc. | Macroblock cache |
| CN101345026A (en) * | 2007-07-10 | 2009-01-14 | 联詠科技股份有限公司 | Frame data buffering apparatus and related frame data acquisition method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103581504B (en) | 2017-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8644381B2 (en) | Apparatus for reference picture resampling generation and method thereof and video decoding system using the same | |
| US9167260B2 (en) | Apparatus and method for video processing | |
| US8619862B2 (en) | Method and device for generating an image data stream, method and device for reconstructing a current image from an image data stream, image data stream and storage medium carrying an image data stream | |
| RU2009122983A (en) | CODING METHOD AND VIDEO DECODING METHOD, DEVICES FOR THEIR IMPLEMENTATION, PROGRAMS FOR THEIR IMPLEMENTATION AND MEMORY CARRIERS STORING PROGRAMS | |
| EP2942965A3 (en) | Moving image prediction encoding device, moving image prediction encoding method, moving image prediction encoding program, moving image prediction decoding device, moving image prediction decoding method, and moving image prediction decoding program | |
| CN112954233B (en) | Video synthesis system and method based on GPU | |
| CN107820091B (en) | Picture processing method and system and picture processing equipment | |
| CN101193306A (en) | Motion vector detecting apparatus and motion vector detecting method | |
| CN102113327B (en) | Image encoding device, method, and integrated circuit | |
| WO2022095797A1 (en) | Image compression method and apparatus, and intelligent terminal and computer-readable storage medium | |
| US20090168893A1 (en) | System, method and device for processing macroblock video data | |
| CN1214648C (en) | Method and apparatus for performing motion compensation in a texture mapping engine | |
| US8687083B2 (en) | High resolution digital image capturing apparatus and reference pixel memory storage space configuration method | |
| TWI601075B (en) | Motion compensation image processing apparatus and image processing method | |
| CN103079072A (en) | Inter-frame prediction method, encoding equipment and decoding equipment | |
| CN113645448A (en) | Video decoding method and device suitable for command scheduling | |
| FI2869577T3 (en) | Dynamic image predictive decoding device, method, and program | |
| CN103581504A (en) | Motion compensated image processing device and image processing method | |
| CN110891195A (en) | Method, device and equipment for generating screen image and storage medium | |
| CN103327340B (en) | A kind of integer searches method and device | |
| CN103248885B (en) | Intra-frame image prediction decoding method and Video Codec | |
| CN105791819A (en) | Frame compression method for image and decompression method and device for image | |
| US20040218670A1 (en) | Method and apparatus for reducing the bandwidth required for transmitting video data for display | |
| US20080273595A1 (en) | Apparatus and related method for processing macroblock units by utilizing buffer devices having different data accessing speeds | |
| CN101990102B (en) | Data reading method and device on embedded equipment in video decoding process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170620 Termination date: 20190720 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |