Digital image interpolation method
AssigneePrimax Electronics Ltd.
InventorsYan-Ru Jiang, Po-Chiang Yeh
Life of this patent
Estimating the term…
Abstract
A digital image interpolation method is disclosed which a projection point (x, y) of the original image corresponding to a pixel of the transformed image first is determined. According to neighboring pixels (m, n) of the projection point, Gauss weights Wi then are determined. Lastly, according to the Gauss weights Wi and the corresponding color value Vi of the neighboring pixels, a color value f (x, y) of the projection point is computed. Via an interpolation method using Gauss weight, the computation time is reduced while qualitatively improving the image processing.
Claims11
A digital image interpolation method, comprising:
- determining a projection point (x, y) of an original image corresponding to a pixel of a transformed image;
- according to the coordinates (m, n) of neighboring pixels of the projection point, determining a plurality of Gauss weights Wi associated to each neighboring pixel;
- and
- according to the Gauss weight Wi and a color value Vi of each neighboring pixel, computing a color value f (x, y) of the projection point;
- wherein determining a plurality of Gauss weights Wi associated to each neighboring pixel according to the coordinates (m, n) of the neighboring pixels of the projection point is performed by computing the following formula:
- Wi=e−r√{square root over ((x−m)2+(y−n)2)},
- wherein r is a coefficient calculated from the relation:
- Σ(i=1, 4) Wi=1.
A digital image interpolation method, comprising:
- determining a projection point (x, y) of an original image corresponding to a pixel of a transformed image;
- according to the coordinates (m, n) of neighboring pixels of the projection point, determining a plurality of Gauss weights Wi associated to each neighboring pixel;
- and
- according to the Gauss weight Wi and a color value Vi of each neighboring pixel, computing a color value f (x, y) of the projection point;
- wherein determining a plurality of Gauss weights Wi associated to each neighboring pixel according to the coordinates (m, n) of the neighboring pixels of the projection point is performed by checking in a weight table, and the weight table is established via a method comprising:
- determining a block area between holes of the neighboring pixels, and attributing a characteristic point in the block area;
- according to the coordinates (m, n) of the neighboring pixels, computing the corresponding Gauss weight;
- and
- recording the Gauss weight associated to the block are.
Dependencies are inferred from the claim text.