摘要 |
<p>Histograms of large volume data, e.g. image data, are calculated in at least two phases under the control of a computer (1), using a first histogram storage (29) with a wast address space and short word length and a second histogram storage (25) with a limited address space but a long word length. In one or several first phases, a preliminary histogram is generated in the first histogram storage; simultaneously, possible overflow items are identified using the second histogram storage. In a final phase, the overflow items are processed in the second histogram storage to update the preliminary histogram in the first histogram storage.</p> |