Motion estimation and motion-compensated interpolition
AssigneeU.S. Philips Corporation
InventorsGerard De Haan, Robert J. Schutten, Anna Pelagotti
Life of this patent
Estimating the term…
Abstract
In a method of estimating motion, at least two motion parameter sets are generated (PE1-PEn) from input video data (n, n−1), a motion parameter set being a set of parameters describing motion in an image, by means of which motion parameter set motion vectors can be calculated. One motion parameter set indicates a zero velocity for all image parts in an image, and each motion parameter set has corresponding local match errors. Output motion data are determined from the input video data (n, n−1) in dependence on the at least two motion parameter sets, wherein the importance of each motion parameter set in calculating the output motion data depends on the motion parameter sets' local match errors.
Claims16
A method of estimating motion in video data, the method comprising the steps:
- furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- determining output motion data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output motion data depends on the motion parameter sets' local match errors,
- wherein said step of furnishing at least two motion parameter sets includes the steps:
- furnishing at least one previously calculated motion parameter set as at least one first prediction parameter vector;
- adding at least one update parameter vector to said at least one first prediction parameter vector to form further prediction parameter vectors;
- and
- selecting one of said first and further prediction parameter vectors.
A method of estimating motion in video data, the method comprising the steps:
- furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- determining output motion data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output motion data depends on the motion parameter sets' local match errors,
- wherein said step of furnishing at least two motion parameter sets includes the step:
- determining, for each motion parameter set other than the zero velocity set, an adjusted motion parameter set in dependence on global match errors, said global match errors being calculated for image parts in accordance with weights assigned to the image parts.
A method of motion-compensating video data, the method comprising the steps:
- furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- interpolating output video data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output video data depends on the motion parameter sets' local match errors,
- wherein said step of furnishing at least two motion parameter sets includes the steps:
- furnishing at least one previously calculated motion parameter set as at least one first prediction parameter vector;
- adding at least one update parameter vector to said at least one first prediction parameter vector to form further prediction parameter vectors;
- and
- selecting one of said first and further prediction parameter vectors.
A method of motion-compensating video data, the method comprising the steps:
- furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- interpolating output video data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output video data depends on the motion parameter sets' local match errors,
- wherein said interpolating step supplies:
- a median of a corresponding pixel in a previous field, a corresponding pixel in a next field, and a motion-compensated average from both said previous and next fields, if the match error of the motion vector used for calculating said motion-compensated average exceeds the match error of the zero motion vector, or
- a median of a motion-compensated pixel from the previous field, a motion-compensated pixel from the next field, and a non-motion-compensated average from both said previous and next fields, if the match error of the motion vector used for furnishing said motion-compensated pixels falls below the match error of the zero motion vector.
A device for estimating motion in video data, the device comprising:
- means for furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- means for determining output motion data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output motion data depends on the motion parameter sets' local match errors,
- wherein said means for furnishing at least two motion parameter sets includes:
- means for furnishing at least one previously calculated motion parameter set as at least one first prediction parameter vector;
- means for adding at least one update parameter vector to said at least one first prediction parameter vector to form further prediction parameter vectors;
- and
- means for selecting one of said first and further prediction parameter vectors.
A device for motion-compensating video data, the device comprising:
- means for furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- means for interpolating output video data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output video data depends on the motion parameter sets' local match errors,
- wherein said interpolating means supplies:
- a median of a corresponding pixel in a previous field, a corresponding pixel in a next field, and a motion-compensated average from both said previous and next fields, if the match error of the motion vector used for calculating said motion-compensated average exceeds the match error of the zero motion vector, or
- a median of a motion-compensated pixel from the previous field, a motion-compensated pixel from the next field, and a non-motion-compensated average from both said previous and next fields, if the match error of the motion vector used for furnishing said motion-compensated pixels falls below the match error of the zero motion vector.
A method of motion-compensating video data, the method comprising the steps:
- furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- interpolating output video data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output video data depends on the motion parameter sets' local match errors,
- wherein said step of furnishing at least two motion parameter sets includes the step:
- determining, for each motion parameter set other than the zero velocity set, an adjusted motion parameter set in dependence on global match errors, said global match errors being calculated for image parts in accordance with weights assigned to the image parts.
A device for estimating motion in video data, the device comprising:
- means for furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- means for determining output motion data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output motion data depends on the motion parameter sets' local match errors,
- wherein said means for furnishing at least two motion parameter sets comprises:
- means for determining, for each motion parameter set other than the zero velocity set, an adjusted motion parameter set in dependence on global match errors, said global match errors being calculated for image parts in accordance with weights assigned to the image parts.
A device for motion-compensating video data, the device comprising:
- means for furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- means for interpolating output video data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output video data depends on the motion parameter sets' local match errors,
- wherein said means for furnishing at least two motion parameter sets comprises:
- means for determining, for each motion parameter set other than the zero velocity set, an adjusted motion parameter set in dependence on global match errors, said global match errors being calculated for image parts in accordance with weights assigned to the image parts.
A device for motion-compensating video data, the device comprising:
- means for furnishing at least two motion parameter sets from input video data, one motion parameter set indicating a zero velocity for all image parts in an image, and each motion parameter set having corresponding local match errors for all image parts in an image, a motion parameter set being a set of parameters describing motion in an image, said motion parameter set being used to calculate motion vectors;
- and
- means for interpolating output video data from said input video data in dependence on said at least two motion parameter sets, wherein the importance of each motion parameter set in calculating said output video data depends on the motion parameter sets' local match errors,
- wherein said means for furnishing at least two motion parameter sets comprises:
- means for furnishing at least one previously calculated motion parameter set as at least one first prediction parameter vector;
- means for adding at least one update parameter vector to said at least one first prediction parameter vector to form further prediction parameter vectors;
- and
- means for selecting one of said first and further prediction parameter vectors.
Dependencies are inferred from the claim text.