| 摘要 |
Points of interest which characterise an image (100) are identified by dividing the image into tiles 301-304 consisting of adjacent pixels; finding within each tile 303 the position of a pixel 304 with an extremum (i.e. maximum or minimum) value and ascribing that extremal value to the tile; identifying a tile with an extremal value which is more extreme than that of all adjacent tiles; and selecting as a point of interest the position 305,306 within the image of the pixel with the extremum value in that identified tile. The image could be a frame of video data, where the pixel values relate to colour or luminance. |
