In addition, a few approaches have been developed dealing with
the reconstruction problem based either on a set of 3D features
which can be possibly found in a technical drawing (e.g.
revolutions and extrusions) or on information coming from
additional views, like cross-sections. The works which confront
the reconstruction problem for curvilinear objects, however,
usually present quite ‘‘tricky’’ approaches, generally involving a
considerable amount of user interaction [8,9].
The most significant limitation to the approaches cited so far is
that, as already mentioned, in all the cases the reconstruction is
based on ‘‘exact’’ vector drawings and is limited to specific kinds of
geometries.