2-D Shape Blending based on Visual Feature Decomposition

Wenwu Yang, Jieqing Feng, Xiaogang Jin, Qunsheng Peng, A. Robin Forrest

Research output: Contribution to conferencePaper


2-D polygon shape blending has been widely used in 2-D character animation, pattern matching, and geometric modeling. Previous algorithms mostly tend to use the polygon’s geometric elements such as edge, angle, area and skeleton, to associate the regions on the two shapes which look alike. However they ignore the visual features of two polygonal shapes. This paper presents a new 2-D polygonal shape blending method based on the correspondence of visual features. By decomposing the two shapes into sub-shape pairs which have similar visual features compatibly, the sub-shape in the source polygon can be smoothly transformed into the corresponding sub-shape in the destination polygon. Since the feature-decomposition vertices are introduced and our blending algorithm is so simple , the user can edit the visual correspondence of features flexibly and intuitively in real time. Experimental results show that the source shape can be transformed into the destination shape as prescribed by the user whilst keeping feature correspondence.
Original languageEnglish
Number of pages8
Publication statusPublished - Jul 2004
Event17th International Conference on Computer Animation and Social Agents - Geneva, Switzerland
Duration: 7 Jul 20049 Jul 2004


Conference17th International Conference on Computer Animation and Social Agents

Cite this