Detection and proper interpolation of interlaced moving areas for MPEG decoding with embedded resizing
View Patent ↗The present invention is directed to MPEG decoding with embedded resizing that includes detecting and interpolating moving areas in interlaced video in order to eliminate irregularities in the output video. According to the present invention, the decoding includes down scaling a motion vector. Also, detecting a moving area in an interlaced video reference frame and modifying the reduced resolution motion vector. Further, retrieving pixel values from the interlaced video reference frame according to the modified motion vector.
1. A method for decoding interlaced video frames, the method comprising the steps of:
down scaling a motion vector to produce a reduced resolution motion vector;
detecting a moving area in an interlaced video reference frame;
modifying the reduced resolution motion vector to produce a modified motion vector; and
retrieving pixel values from the interlaced video reference frame according to the modified motion vector;
wherein the detecting the moving area in the interlaced video reference frame includes:
calculating a difference between two fields at an area of the interlaced video reference frame corresponding to the motion vector; and
comparing the difference between the two fields to a predetermined threshold.
2. The method of claim 1 , wherein the reduced resolution motion vector is modified so that pixels from each field are given an equal weight while performing an interpolation.
3. The method of claim 1 , wherein the reduced resolution motion vector is modified so that the pixel values are retrieved in a correct field order.
4. The method of claim 1 , wherein a vertical component of the reduced resolution motion vector is modified by changing an odd integer to a nearest even integer.
5. The method of claim 1 , wherein a vertical component of the reduced resolution motion vector is modified by changing a “¼” pixel position to a “½” pixel position.
6. The method of claim 1 , wherein a vertical component of the reduced resolution motion vector is modified by changing a “¾” pixel position to a “½” pixel position.
7. The method of claim 1 , which further includes determining if frame motion compensation was performed on each macro block of a video sequence.
8. The method of claim 1 , wherein the difference between the two fields is calculated by accumulating the difference between adjacent pixels of the two fields in at least one column of the area of the interlaced video reference frame.
9. A method for decoding interlaced video frames, the method comprising the steps of:
down scaling a motion vector to produce a reduced resolution motion vector;
detecting a moving area in an interlaced video reference frame;
modifying the reduced resolution motion vector to produce a modified motion vector; and
retrieving pixel values from the interlaced video reference frame according to the modified motion vector;
wherein a vertical component of the reduced resolution motion vector is modified by changing a “½” pixel position to the nearest odd integer position.
10. A memory medium including code for decoding interlaced video frames, the code comprising;
a code for down scaling a motion vector to produce a reduced resolution motion vector;
a code for detecting a moving area in an interlaced video reference frame;
a code for modifying the reduced resolution motion vector to produce a modified motion vector; and
a code for retrieving pixel values from the interlaced video reference frame according to the modified motion vector;
wherein the code for detecting the moving area in the interlaced video reference frame includes:
a code for calculating a difference between two fields at an area of the interlaced video reference frame corresponding to the motion vector; and
a code for comparing the difference between the two fields to a predetermined threshold.
11. A decoder, comprising:
a first path for producing residual error frames;
a second path for producing motion compensated residual frames; and
an adder for combining the residual error frames and the motion compensated residual frames;
wherein the second path includes:
a down scaler for producing a reduced resolution motion vector;
a motion compensation unit for detecting a moving area in an interlaced video reference frame, producing a modified motion vector and retrieving pixel values from the interlaced video reference frame according to the modified motion vector;
wherein the detecting the moving area in the interlaced video reference frame includes:
calculating a difference between two fields at an area of the interlaced video reference frame corresponding to the motion vector; and
comparing the difference between the two fields to a predetermined threshold.