Motion vector correction device and method and video signal processing apparatus and method
AssigneeJVC KENWOOD Corporation
InventorsTomoyuki Shishido, Hiroshi Noguchi, Hideki Aiba
Life of this patent
Estimating the term…
Abstract
A motion vector detector detects, using pixel data in at least two real frames in an input video signal, a motion vector MV1 necessary for generating interpolated pixel data forming an interpolated frame to be inserted between the two real frames. A motion vector corrector corrects the motion vector MV1 to decrease the magnitude of motion vector MV1 when the magnitude of motion vector MV1 exceeds a predetermined threshold, and outputs it as a motion vector MV3.
Claims16
A motion vector correction device comprising:
- a circuitry that
- detects, using pixel data in at least two real frames in an input video signal, a first motion vector necessary for generating interpolated pixel data forming an interpolated frame to be inserted between the two real frames, and
- corrects the first motion vector to decrease the magnitude of first motion vector when the magnitude of first motion vector exceeds a first threshold,
- wherein
- when the magnitude of first motion vector exceeds the first threshold, the circuitry decreases the magnitude of first motion vector within a range where the magnitude of first motion vector is larger than the first threshold such that a decrement amount of the magnitude of first motion vector increases as a value between the magnitude of first motion vector and the first threshold increases, and outputs the result as a second motion vector,
- when the magnitude of second motion vector exceeds a second threshold, the circuitry keeps the magnitude of second motion vector unchanged from the second threshold to a third threshold larger than the second threshold, and outputs the result as a third motion vector, and
- when the magnitude of second motion vector exceeds the third threshold, the circuitry decreases the magnitude of second motion vector within a range where the magnitude of second motion vector is smaller than the second threshold such that a decrement amount of the magnitude of second motion vector increases as a value between the magnitude of second motion vector and the third threshold increases, and outputs the result as a third motion vector.
A motion vector correction device comprising:
- a circuitry that
- detects, using pixel data in at least two real frames in an input video signal, a first motion vector necessary for generating interpolated pixel data forming an interpolated frame to be inserted between the two real frames, and
- corrects the first motion vector to decrease the magnitude of first motion vector when the magnitude of first motion vector exceeds a first threshold,
- wherein
- when the magnitude of first motion vector exceeds the first threshold, the circuitry decreases the magnitude of first motion vector within a range where the magnitude of first motion vector is larger than the first threshold such that a decrement amount of the magnitude of first motion vector increases as a value between the magnitude of first motion vector and the first threshold increases, and outputs the result as a second motion vector,
- when the magnitude of second motion vector exceeds a second threshold, the circuitry keeps the magnitude of second motion vector unchanged from the second threshold to a third threshold larger than the second threshold, and outputs the result as a third motion vector, and
- when the magnitude of second motion vector exceeds the third threshold, the circuitry decreases the magnitude of second motion vector within a range where the magnitude of second motion vector is smaller than the second threshold such that a decrement amount of the magnitude of second motion vector increases as a value between the magnitude of second motion vector and the third threshold increases, and outputs the result as a third motion vector, and
- wherein the circuitry calculates the second motion vector according to the following calculus equation
- the magnitude of first motion vector)×m/n+the first threshold)×(n−m)/n,
- where variables m and n are natural numbers which have a relation of m<n.
A motion vector correction method comprising:
- a motion vector detection step for, using pixel data in at least two real frames in an input video signal, detecting a first motion vector necessary for generating interpolated pixel data forming an interpolated frame to be inserted between the two real frames;
- and
- a motion vector correction step for correcting the first motion vector to decrease the magnitude of first motion vector when the magnitude of first motion vector exceeds a first threshold,
- wherein the motion vector correction step includes:
- a motion vector value suppressing step for decreasing, when the magnitude of first motion vector exceeds the first threshold, the magnitude of first motion vector within a range where the magnitude of first motion vector is larger than the first threshold such that a decrement amount of the magnitude of first motion vector increases as a value between the magnitude of first motion vector and the first threshold increases, and outputting the result as a second motion vector;
- and
- a motion vector value converting step for
- keeping, when the magnitude of second motion vector exceeds a second threshold, the magnitude of second motion vector unchanged from the second threshold to a third threshold larger than the second threshold, and outputting the result as a third motion vector, and
- decreasing, when the magnitude of second motion vector exceeds the third threshold, the magnitude of second motion vector within a range where the magnitude of second motion vector is smaller than the second threshold such that a decrement amount of the magnitude of second motion vector increases as a value between the magnitude of second motion vector and the third threshold increases, and outputting the result as a third motion vector.
Dependencies are inferred from the claim text.