Katie Boundary
9th November 2016, 01:12
I'd like feedback on the following approach to bob-deinterlacing:
Check the values of the pixels directly above and below the current pixel, as well as the pixels in the same spot in the next and previous fields. If the previous pixel has R, B, and G values between those of the above and below pixels, it "passes". Repeat for the next pixel. If neither [p] nor [n] passes, then ignore them and perform linear spatial interpolation (iBob). If both pass, then ignore [a] and [b], and perform linear temporal interpolation. If [p] passes but [n] doesn't, then simply copy [p]'s values to [c]. If [n] passes but [p] doesn't, then simply copy [n]'s values to [c].
(obviously would need to be adjusted for edge/corner pixels and first/last frames)
?
Check the values of the pixels directly above and below the current pixel, as well as the pixels in the same spot in the next and previous fields. If the previous pixel has R, B, and G values between those of the above and below pixels, it "passes". Repeat for the next pixel. If neither [p] nor [n] passes, then ignore them and perform linear spatial interpolation (iBob). If both pass, then ignore [a] and [b], and perform linear temporal interpolation. If [p] passes but [n] doesn't, then simply copy [p]'s values to [c]. If [n] passes but [p] doesn't, then simply copy [n]'s values to [c].
(obviously would need to be adjusted for edge/corner pixels and first/last frames)
?