Motion object video on film detection and adaptive de-interlace method based on fuzzy logic
View Patent ↗A method and system for performing fuzzy logic based de-interlacing on film source fields that might be mixed with video on film. An embodiment of the invention comprises an adaptive de-interlacer by weighing between merge operation and interpolation operation in the case of occurring video on film motion object. A weighing factor is generated from video on film pattern based on fuzzy logic inference rules. This weighing factor specifies the weighting between merging and interpolating in assigning the pixel values of the progressive display output.
1. A computer-implemented method, comprising:
assessing a likelihood of having a video on film (VOF) object edge along one of n pre-defined directions, the assessing comprising:
generating n sets of reference quantities, wherein each reference quantity is associated with a corresponding one of n pre-defined directions; and
generating two additional reference quantities R[1] and R[2] by applying classical logic to the n sets of reference quantities;
determining a weighting factor k by analyzing R[1] and R[2] using fuzzy logic interference rules; and
performing de-interlacing at a current processing pixel based at least in part on the weighting factor k.
2. The computer-implemented method of claim 1 , wherein the n pre-defined directions intercept approximately at the current processing pixel.
3. The computer-implemented method of claim 1 , wherein one of the reference quantities comprises a pixel value difference between two inter-field pixels that each neighbor the current processing pixel and lie along a pre-defined direction through the current processing pixel.
4. The computer-implemented method of claim 1 , wherein one of the reference quantities comprises a pixel value difference between the current processing pixel and an intra-field pixel that neighbors the current processing pixel.
5. The computer-implemented method of claim 1 , wherein one of the reference quantities comprises a pixel value difference between two intra-field pixels that each neighbor the current processing pixel and lie along a pre-defined direction through the current processing pixel.
6. The computer-implemented method of claim 1 , wherein R[1] comprises an intra-field reference quantity that quantifies a maximum dissimilarity between a pair of intra-field pixels of the current processing pixel lying along the selected direction.
7. The computer-implemented method of claim 1 , wherein R[2] comprises an inter-field reference quantity that quantifies a minimum similarity between a pair of inter-field pixels of the current processing pixel lying along the selected direction.
8. The computer-implemented method of claim 1 , further comprising receiving the weighting factor k by an adaptive de-interlacing unit.
9. The computer-implemented method of claim 8 , wherein the adaptive de-interlacing unit performs the de-interlacing.
10. A computer-implemented method, comprising:
storing within each of a plurality of buffers a corresponding one of a plurality of video fields;
receiving at an interpolation unit the plurality of video fields stored within the plurality of buffers;
generating a weighting factor, wherein the weighting factor represents an assessed likelihood of having a video on film object edge through a current processing pixel of the plurality of video fields.
11. The computer-implemented method of claim 10 , further comprising performing a de-interlacing operation based on the assessed likelihood as represented by the weighting factor.
12. The computer-implemented method of claim 10 , wherein the generating comprises using fuzzy logic inference rules.
13. The computer-implemented method of claim 10 , wherein the de-interlacing operation comprises a merging de-interlacing operation.
14. The computer-implemented method of claim 10 , wherein the de-interlacing operation comprises an interpolation de-interlacing operation.
15. The computer-implemented method of claim 14 , wherein the interpolation de-interlacing operation comprises at least one of an intra-field interpolation operation and an inter-field interpolation operation.
16. A computer-implemented method, comprising:
determining an ability of multiple de-interlacing operations to de-interlace image field data using fuzzy logic, wherein the image field data comprises a pixel within an image field;
de-interlacing the image field data according to the fuzzy logic determination;
determining a coefficient according to at least one fuzzy logic inference rule; and
blending the de-interlaced image field data from the multiple de-interlacing operations according to the fuzzy logic determination and according to the coefficient.
17. The computer-implemented method of claim 16 , wherein de-interlacing the image field data comprises de-interlacing the image field data according to the multiple de-interlacing operations.
18. The computer-implemented method of claim 16 , further comprising identifying an image transition through the pixel.
19. The computer-implemented method of claim 18 , wherein the image transition corresponds to an edge of a picture-in-picture application.