| 摘要 |
Edge pixels in an image are detected by replicating at least part of a first image to create at least one replicated image which is then spatially transform by a distance of one or more pixels relative to the first image. An edge pixel matrix indicating the presence of edge pixels in the at least part of the first image is then produced wherein the matrix comprises a plurality of matrix elements, each matrix element corresponding to a pixel in the at least part of the first image. A value for each element of the edge pixel matrix is then generated in accordance with a logical exclusive OR (XOR) combination of a same attribute of a pixel in the first image with one corresponding pixel from at least part of the transformed image, wherein said corresponding pixel from the transformed image is of a matching position to the pixel from the first image, the logical combination resulting in a first value only if both the pixels have the same attribute. |
