Patent US6891891B2 successfully refreshed from Google Patents.
US6891891B2
Expired / Lapsed
US
⏱️ 0.0 years remaining
Motion estimation process and system
Assignee: STMicroelectronics SRL
•
Filing: 2001-05-04
•
Granted: 2005-05-10
•
Expires: 2021-05-04
Abstract
Abstract A motion estimation process in video signals organized in successive frames divided into macroblocks that is carried out by the identification of motion vectors. In a first identification phase, starting from a current motion vector, a best motion vector predictor is identified, chosen from a set of candidates. The best predictor thus identified is then subjected to a second refining phase. The aforesaid set of candidates is identified by selecting vectors belonging to macroblocks close to the current vector within the current frame and the preceding frame. Preferably, the refining phase comprises the definition of a grid of n points centered on the central position to which the best motion vector points and the distance of the points of the grid from the center is defined as a function of the matching error typically consisting of an SAD function, defined in the first identification phase. Application to the IPB and APM operating modes of the H.263+ video standard is envisaged.
Inventors
- 👤 Danilo Pau
- 👤 Emiliano Piccinelli
- 👤 Fabrizio Rovati
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 10:15
CPC Classifications
H; H04; H04N; H04N5/00; H04N5/14; H04N5/144; H04N5/145; H04N19/00; H04N19/50; H04N19/503; H04N19/51; H04N19/56
Claims Analysis
Total Claims: 39
Claim 1
Independent Claim
130 words
A process for motion estimation in video signals organized in successive frames divided into macroblocks, the process comprising: a first identification phase in which, starting from a current motion vector, a best motion vector predictor is identified within a set of candidates, and a second phase of refining the best motion vector predictor thus identified; the set of candidates formed from vectors belonging to macroblocks association with the current vector within the current frame and the preceding frame, the refining phase comprising the operation of forming a grid of n points centered on the central position to which a best mode vector points, the distance of the points of the grid from the center being defined as a linear function of a matching error defined in the first identification phase.
Claim 2
Dependent Claim
130 words
A process for motion estimation in video signals organized in successive frames divided into macroblocks, the process comprising: a first identification phase in which, starting from a current motion vector, a best motion vector predictor is identified within a set of candidates, and a second phase of refining the best motion vector predictor thus identified; the set of candidates formed from vectors belonging to macroblocks association with the current vector within the current frame and the preceding frame, the refining phase comprising the operation of forming a grid of n points centered on the central position to which a best mode vector points, the distance of the points of the grid from the center being defined as a linear function of a matching error defined in the first identification phase.
Claim 3
Dependent Claim
23 words
The process of claim 1 , wherein the set includes, in the preceding frame, the vector homologous to the said current motion vector.
Claim 4
Dependent Claim
45 words
The process of claim 1 , compromising the operation of scanning the macroblocks in lexicographic order within each frame, as a result of which the set includes, within the current frame, vectors belonging to macroblocks located above and to the left of the current macroblock.
Claim 5
Dependent Claim
59 words
The process of claim 1 , wherein the set includes vectors belonging to at least one of the macroblocks chosen from the group consisting of: in the current frame, macroblocks located to the left of and above the current macroblock, and in the preceding frame, the homologous macroblock and the macroblocks located to the left of the current macroblock.
Claim 6
Dependent Claim
25 words
The process of claim 1 , wherein the best predictor is identified, within the set, as the predictor that minimizes a residual error measurement function.
Claim 7
Dependent Claim
13 words
The process of claim 5 , wherein the function is the SAD function.
Claim 8
Dependent Claim
13 words
The process of claim 5 , wherein the function is the SSD function.
Claim 9
Dependent Claim
21 words
The process of claim 1 , wherein the distance is defined as the product of a coefficient and the error function.
Claim 10
Dependent Claim
14 words
The process of claim 1 , wherein the error function is the SAD function.
Claim 11
Dependent Claim
33 words
The process of claim 1 , comprising the operation of selecting as the motion vector for the current macroblock, within the said grid, the motion vector that minimizes a further residual error function.
Claim 12
Dependent Claim
16 words
The process of claim 10 , wherein the further residual error function is the SAD function.
Claim 13
Dependent Claim
19 words
The process of claim 1 , comprising the operation of selectively amplifying the grid by a given amplification factor.
Claim 14
Dependent Claim
42 words
The process of claim 12 , comprising the operation of monitoring the quality of the predictor as a function of a given error function, and wherein the grid is amplified when the predictor has an error function greater than the given value.
Claim 15
Dependent Claim
30 words
The process of claim 13 , wherein the grid amplification factor is calculated as a function of the analysis of the distribution of the error function for some sample sequences.
Claim 16
Dependent Claim
53 words
The process of claim 1 , applied to the IPB operating mode of the H.263+ video standard for the processing of pairs of PB frames, comprising the operation of selecting the temporal predictors of the B frame of a BP pair before the temporal predictors of the P frame of the BP pair.
Claim 17
Dependent Claim
83 words
The process of claim 1 , applied to the APM operating mode of the H.263+ video standard, in which the frame of a macroblock is associated with a plurality of vectors, preferably four in number, associated with a corresponding plurality of further blocks constituting the macroblock to which the previously found motion vector points, in which the first identification phase is applied to the macroblock, while the second phase is applied to the blocks of the plurality to identify the plurality of vectors.
Claim 18
Independent Claim
148 words
A system for motion estimation in video signals organized in successive frames divided into macroblocks by means of the identification of motion vectors, the system comprising: a motion estimation block configured to implement a first identification phase in which, starting from a current motion vector, a best motion vector predictor is identified within a set of candidates, and a second phase of refining the best predictor thus identified; the said set of candidates being formed from vectors belonging to macroblocks close to the current vector within the current frame and the preceding frame, the estimation block configured to implement the refining phase by forming a grid of n points centered on the central position to which the best mode vector points, the distance of the points of the grid from the center defined as the function, preferably linear, of the matching error defined in the first identification phase.
Claim 19
Independent Claim
148 words
A system for motion estimation in video signals organized in successive frames divided into macroblocks by means of the identification of motion vectors, the system comprising: a motion estimation block configured to implement a first identification phase in which, starting from a current motion vector, a best motion vector predictor is identified within a set of candidates, and a second phase of refining the best predictor thus identified; the said set of candidates being formed from vectors belonging to macroblocks close to the current vector within the current frame and the preceding frame, the estimation block configured to implement the refining phase by forming a grid of n points centered on the central position to which the best mode vector points, the distance of the points of the grid from the center defined as the function, preferably linear, of the matching error defined in the first identification phase.
Claim 20
Dependent Claim
29 words
The system of claim 17 , wherein the estimation block is configured to include in the set in the preceding frame the vector homologous to the current motion vector.
Claim 21
Dependent Claim
49 words
The system of claim 17 , wherein the motion estimation block is configured to scan the macroblocks in lexicographic order within each frame, as a result of which the set includes, within the current frame, vectors belonging to macroblocks located above and to the left of the current macroblock.
Claim 22
Dependent Claim
62 words
The system of claim 17 , wherein the estimation block includes in the set vectors belonging to at least one of the macroblocks chosen from the group comprising: in the current frame, macroblocks located to the left of and above the current macroblock, and in the preceding frame, the homologous macroblock and the macroblocks located to the left of the current macroblock.
Claim 23
Dependent Claim
27 words
The system of claim 17 , wherein the estimation block selects the best predictor, within the set, as the predictor that minimizes a residual error measurement function.
Claim 24
Dependent Claim
13 words
The system of claim 21 , wherein the function is the SAD function.
Claim 25
Dependent Claim
13 words
The system of claim 21 , wherein the function is the SSD function.
Claim 26
Dependent Claim
21 words
The system of claim 17 , wherein the distance is defined as the product of a coefficient and the error function.
Claim 27
Dependent Claim
14 words
The system of claim 17 , wherein the error function is the SAD function.
Claim 28
Dependent Claim
31 words
The system of claim 17 , wherein the estimation block selects, as the motion vector for the current macroblock, within the grid, the vector that minimizes a further residual error function.
Claim 29
Dependent Claim
16 words
The system of claim 10 , wherein the further residual error function is the SAD function.
Claim 30
Dependent Claim
22 words
The system of claim 17 , wherein the estimation block is capable of selectively amplifying the grid by a given amplification factor.
Claim 31
Dependent Claim
41 words
The system according to claim 28 , wherein the estimation block monitors the quality of the predictor as a function of a given error function, and amplifies the grid when the predictor has an error function greater than the given value.
Claim 32
Dependent Claim
32 words
The system of claim 28 , wherein the estimation block calculates the grid amplification factor as a function of the analysis of the distribution of the error function for some sample sequences.
Claim 33
Dependent Claim
57 words
The system of claim 17 , configured to operate according to the IPB operating mode of the H.263+ video standard for the processing of pairs of PB frames, in which the estimation block selects the temporal predictors of the B frame of a BP pair before the temporal predictors of the P frame of the BP pair.
Claim 34
Dependent Claim
89 words
The system of claim 17 , configured to operate according to the APM operating mode of the H.263+ video standard, in which the estimation block associates the frame of a macroblock with a plurality of vectors, preferably four in number, associated with a corresponding plurality of further blocks constituting the macroblock to which the previously found motion vector points, in which the estimation block applies, respectively, the first identification phase to the macroblock, and the second phase to the blocks of the plurality to identify the plurality of vectors.
Claim 35
Dependent Claim
137 words
A motion estimation process, comprising: identifying a best motion vector predictor as the predictor that minimizes a residual error measurement function within a set of candidates formed from vectors belonging to at least one macroblock chosen from a group of macroblocks in the current frame of macroblocks located to the left of and above the current macroblock, and in the preceding frame the homologous macroblock and the macroblocks located to the left of the current macroblock; and refining the best motion vector predictor so identified, comprising the operation of forming a grid of n points centered on a central position to which a best mode vector points, including defining the distance of the best mode vector pointing to the center as a linear function of a matching error previously defined in identifying the best motion vector predictor.
Claim 36
Dependent Claim
137 words
A motion estimation process, comprising: identifying a best motion vector predictor as the predictor that minimizes a residual error measurement function within a set of candidates formed from vectors belonging to at least one macroblock chosen from a group of macroblocks in the current frame of macroblocks located to the left of and above the current macroblock, and in the preceding frame the homologous macroblock and the macroblocks located to the left of the current macroblock; and refining the best motion vector predictor so identified, comprising the operation of forming a grid of n points centered on a central position to which a best mode vector points, including defining the distance of the best mode vector pointing to the center as a linear function of a matching error previously defined in identifying the best motion vector predictor.
Claim 37
Dependent Claim
20 words
The process of claim 33 , wherein the residual error measurement function comprises a SAD (sum of absolute differences) function.
Claim 38
Dependent Claim
20 words
The process of claim 33 , wherein the residual error measurement function comprises a SSD (sum of squared differences) function.
Claim 39
Dependent Claim
23 words
The process of claim 33 , wherein the distance is defined as a product of a coefficient of the residual error measurement function.
References Cited (Backward Citations)
US5786860A
Cited
US5581308A
Cited
US5668608A
Cited
US5745183A
Cited
US6563872B2
Cited
US6014181A
Cited
US6512550B1
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2008-05-09 | 2008-11-08 | 2009-05-10 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2012-05-09 | 2012-11-08 | 2013-05-10 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2016-05-09 | 2016-11-08 | 2017-05-10 | $7700.00 | Window Closed / Lapsed |