| 摘要 |
A method and apparatus for the dc-interlacing of video data are described, wherein pixel interpolation is performed for lines missing in the original interlaced fields. Correlation data is generated for a number of possible interpolations, which is used to calculate a confidence measure relating to this data. Consequently, the correlation data is adjusted on the basis of the confidence measure in order to determine the most accurate interpolation scheme, which is selected and applied for each pixel on the missing line. The figure demonstrates how this is performed in practice. The correlation curve (-) represents the degree of difference from adjacent pixels for a number of possible interpolations. The minima identify the most promising interpolations. A confidence measure is calculated from the correlation data to produce adjustment data (ie 'HappyCurve'). This adjustment data is applied to the correlation data to produce a resultant adjusted correlation curve (o), which identifies a single minimum relating to the interpolation scheme most likely to return an accurate pixel interpolation. Pixel interpolation is subsequently applied according to the selected scheme. |
