Detecting geometric object in digital image file based on reference object information, using relative positions and shapes of sub-objects given be set of transformations

摘要

The method involves dividing the reference object (O) into two or more spatially different sub-objects, and determining a set of transformations indicative of how a sub-object is arranged in terms of its location and shape relative to at least one other sub-object. Individual sub-objects are searched in the digital image file. The search step is repeated until all sub-objects have been found, or a termination criterion has been fulfilled. The found geometric object is verified using the relative positions and shapes of the sub-objects given by the set of transformations.