| 摘要 |
Scanned region type sensitivity logic (110) and scanned region clustering/declustering logic (120) allow the manipulation (36, 37) of regions within a scanned image. A scanned region type sensitivity logic (110) allows the manipulation (36) of region type attributes by allowing a user of a scanner (11) product, through the use of a user interface (13), to adjust the sensitivity (36) of a particular region type in real time, with the results of the adjustment displayed back to the user through the user interface (13). A scanned region clustering/declustering logic (120) allows a user of a scanner (11) product, through the use of a user interface (13), to manipulate (37) the grouping of regions displayed to the user in real time. Both of these forms of scanning logic are predicated on the underlying use of a document analysis technology (17) that stores probabilities ("p-values") or other relative statistics on all plausible region types as part of the definition of the regions. <IMAGE> <IMAGE> |
