基于骨架提取特征的可控渐变算法

提出了一种基于视觉特征的渐变算法,通过图像骨架确立特征点,进而对图形进行同构特征分解。采用可控制子多边形边角插值,通过轮廓多边形渐变控制子多边形渐变,用户也可以通过控制轮廓插值交互控制渐变过程。实验表明:该算法简单有效渐变过程自然合理。

Controllable morphing method using skeleton to extract features

This paper presents a new morphing method based on the visual features.By computing the images’ skeletons，the feature vertices are decided and the shapes’ isomorphic decomposition is built then.A controllable intrinsic interpolation algorithm for local polygons is presented using the interpolation of contour polygon to control the interpolation of local polygons.Users can also control the morphing by changing the interpolation of contour polygon.Experimental results show that this method is easy and robust，and leads to natural results.