| 摘要 |
A method and system for identifying files by calculating a unique table of values for the file. The calculated values in the table constitute a "footprint" of the file. A first embodiment works for any type of files. A second embodiment is particularly useful in identifying image files. In the first embodiment, for each possible value of the bytes that form the file, the offset of the first occurrence of a byte with the particular value, the offset of the last occurrence of a byte with the particular value, and the number of occurrences of the bytes with the particular value are calculated. These values are stored in a table and termed prominence values. The prominence values uniquely identify the file. In a second embodiment, an image file is broken into relatively small cells. The cells with the highest and lowest average values of quantities such as luminance, red chrominance, etc. are determined. The cells with these values are the prominences for the particular image. Vectors which describe the location of the prominence cells are calculated and saved. The values of the various quantities in the prominence cells and the vectors pointing to the location of these cells provide a footprint for the image.
|
