Video de-interlacer using pixel trajectory
A method of de-interlacing interlaced video information including determining functional equations which estimate trajectories of corresponding pixel locations based on statistical information, updating each functional equation with sampled pixel values from the interlaced video information of corresponding pixel locations, and evaluating the functional equations at a time point of a progressive frame and providing corresponding progressive pixel values. A video de-interlace system including a trajectory estimator and a component estimator. The trajectory estimator provides functional equations estimating trajectories of tracked pixel locations based on statistical information. The component estimator receives the functional equations and the interlaced video information and provides progressive pixel values. The component estimator updates each functional equation with sampled pixel values from the interlaced video information, and evaluates the updated functional equations at selected time points to provide progressive frames.
1. A method of de-interlacing interlaced video information, comprising:
determining a plurality of functional equations based on statistical information, wherein each of the plurality of functional equations estimates a corresponding one of a plurality of trajectories, and wherein each of the plurality of trajectories represents how a value of a corresponding one of a plurality of pixel locations changes over time;
updating each of the plurality of functional equations with sampled pixel values from the interlaced video information of corresponding pixel locations; and
evaluating the plurality of functional equations at a time point of a progressive frame and providing a corresponding plurality of progressive pixel values.
2. The method of claim 1 , wherein said determining a plurality of functional equations comprises determining a coefficient for each of at least one Kalman filter.
3. The method of claim 1 , wherein said determining a plurality of functional equations comprises determining a coefficient for each of at least one H-Infinity filter.
4. The method of claim 1 , wherein said determining a plurality of functional equations comprises estimating noise incorporated within the interlaced video information.
5. The method of claim 1 , further comprising determining the statistical information based on the interlaced video information.
6. The method of claim 1 , further comprising:
averaging information of a selected region of the progressive frame and providing averaged information; and
updating the statistical information based on the averaged information.
7. The method of claim 1 , further comprising:
said determining comprising tracking a subset of the pixel locations; and
interpolating a pixel location of the progressive frame that is not tracked based on tracked pixel locations.
8. The method of claim 7 , further comprising:
selecting a plurality of the tracked pixel locations that are correlated with the pixel location that is not tracked; and
wherein said interpolating comprises:
updating a functional equation of each correlated pixel location with sampled pixel values of the pixel location that is not tracked from the interlaced video information;
evaluating the updated functional equations at the time point of the progressive frame to provide a plurality of correlated pixel values; and
combining the plurality of correlated pixel values with a weighting function to provide a pixel value for the pixel location that is not tracked.
9. The method of claim 8 , wherein said selecting comprises selecting a tracked pixel location that is within a predetermined proximity of the pixel location that is not tracked.
10. The method of claim 1 , wherein:
said determining a plurality of functional equations comprises determining an odd set of functional equations for pixel locations within an odd field and determining an even set of functional equations for pixel locations within an even field; and
wherein said updating each of the plurality of functional equations comprises updating each of the odd set of functional equations with sampled pixel values for the odd field and updating each of the even set of functional equations with sampled pixel values for the even field.
11. A video de-interlace system, comprising:
a trajectory estimator having an output providing a plurality of functional equations based on statistical information, wherein each of said plurality of functional equations estimates a corresponding one of a plurality of trajectories, and wherein each of said plurality of trajectories represents how a value of a corresponding one of a plurality of pixel locations changes over time; and
a component estimator having a first input receiving said plurality of functional equations, a second input receiving interlaced video information, and an output providing a plurality of progressive pixel values of a progressive frame;
wherein said component estimator updates each of said plurality of functional equations with corresponding sampled pixel values from said interlaced video information and evaluates said updated plurality of functional equations at a time point of said progressive frame to provide said plurality of progressive pixel values.
12. The video de-interlace system of claim 11 , wherein said trajectory estimator comprises at least one Kalman filter.
13. The video de-interlace system of claim 11 , wherein said trajectory estimator comprises at least one H-Infinity filter.
14. The video de-interlace system of claim 11 , further comprising:
a region averaging control circuit having an input receiving said plurality of progressive pixel values and an output providing statistical adjust information; and
a statistics calculation circuit having an input coupled to said output of said region averaging control circuit and an output providing said statistical information.
15. The video de-interlace system of claim 14 , wherein said statistics calculation circuit has an input receiving said interlaced video information for updating said statistical information.
16. The video de-interlace system of claim 11 , wherein said component estimator selects a plurality of said tracked pixel locations as correlated pixel locations that are correlated with a non-tracked pixel location, updates a functional equation of each said correlated pixel location with sampled pixel values from the interlaced video information corresponding to said non-tracked pixel location, evaluates updated functional equations at a time point of said progressive frame to provide a plurality of correlated pixel values, and combines said correlated pixel values using a weighting function to determine a pixel value for said non-tracked pixel location.
17. The video de-interlace system of claim 16 , wherein said component estimator selects said correlated pixel locations based on relative proximity to said non-tracked pixel location.
18. The video de-interlace system of claim 11 , wherein:
said trajectory estimator comprises:
an odd field trajectory estimator having an output providing a plurality of odd field functional equations estimating a plurality of trajectories of a corresponding plurality of tracked pixel locations within an odd field; and
an even field trajectory estimator having an output providing a plurality of functional equations estimating a plurality of trajectories of a corresponding plurality of tracked pixel locations within an even field; and
wherein said component estimator comprises:
an odd field component estimator having a first input receiving said plurality of odd field functional equations, a second input receiving odd field interlaced video information, and an output providing a plurality of odd field progressive pixel values; and
an even field component estimator having a first input receiving said plurality of even field functional equations, a second input receiving even field interlaced video information, and an output providing a plurality of even field progressive pixel values.
19. A de-interlacer, comprising:
a first trajectory estimator having an input receiving first statistical adjust information and an output providing a plurality of first functional equations for a first field based on said first statistical adjust information, wherein each of said plurality of first functional equations estimates a corresponding one of a plurality of first trajectories of said first field, and wherein each of said plurality of first trajectories represents how a value of a corresponding one of a plurality of first pixel locations of said first field changes over time;
a second trajectory estimator having an input receiving second statistical adjust information and an output providing a plurality of second functional equations for a second field based on said second statistical adjust information, wherein each of said plurality of second functional equations estimates a corresponding one of a plurality of second trajectories of said second field, and wherein each of said plurality of second trajectories represents how a value of a corresponding one of a plurality of second pixel locations of said second field changes over time;
a first component estimator having a first input receiving said plurality of first functional equations, a second input receiving first interlaced video information, and an output providing a first plurality of progressive pixel values;
a second component estimator having a first input receiving said plurality of second functional equations, a second input receiving second interlaced video information, and an output providing a second plurality of progressive pixel values;
a frame buffer having first and second inputs receiving said first and second pluralities of progressive pixel values and an output providing progressive video information;
a region averaging control circuit having an input receiving said progressive video information and an output providing statistical information based on selected regions of said progressive video information; and
a statistics calculations circuit having a first input coupled to said output of said region averaging circuit, a first output providing said first statistical adjust information, and a second output providing said second statistical adjust information.
20. The de-interlacer of claim 19 , wherein said statistics calculations circuit has a second input receiving said first and second interlaced video information.
21. The de-interlacer of claim 19 , wherein said first and second trajectory estimators each comprise a plurality of Kalman filters.
22. The de-interlacer of claim 19 , wherein said first and second trajectory estimators each comprise a plurality of H-Infinity filters.