Size functions, a fairly new class of topological transforms, are used to implement SketchUp, a demonstrative classifier of free-hand drawn sketches. Sketches are drawn without a template and in total freedom. Four different such transforms cooperate to allow recognition in a seven item list of tools. 207 drawings have been examined, with 87.9% of success, 8.7% of rejection, 3.4% of errors. The training set is very small: 66 drawings; none of the drawers was common to the training and test sets.
Sketch up: Towards qualitative shape data management / Collina C.; Ferri M.; Frosini P.; Porcellini E.. - STAMPA. - 1351:(1997), pp. 338-345. (Intervento presentato al convegno 3rd Asian Conference on Computer Vision, ACCV 1998 tenutosi a chn nel 1998) [10.1007/3-540-63930-6_139].
Sketch up: Towards qualitative shape data management
Frosini P.;
1997
Abstract
Size functions, a fairly new class of topological transforms, are used to implement SketchUp, a demonstrative classifier of free-hand drawn sketches. Sketches are drawn without a template and in total freedom. Four different such transforms cooperate to allow recognition in a seven item list of tools. 207 drawings have been examined, with 87.9% of success, 8.7% of rejection, 3.4% of errors. The training set is very small: 66 drawings; none of the drawers was common to the training and test sets.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
