Lecture 7: Intro to Geometry, Splines, and Bezier Curves (75)
atsu-kotani
If our task were to find the best curve that fits to the target polynomial (as in CAD design), I think it would help to have a large number of control points. But as described here, there is a tradeoff; it would often cause problems in interpolation and also some computational complexity, and it is generally better to use low-D curve representations. As briefly mentioned in the earlier slide, I thought NURBS was a better alternative here as it would lower the number of required control points by putting weights on them, without sacrificing the expressibility of the curve..
If our task were to find the best curve that fits to the target polynomial (as in CAD design), I think it would help to have a large number of control points. But as described here, there is a tradeoff; it would often cause problems in interpolation and also some computational complexity, and it is generally better to use low-D curve representations. As briefly mentioned in the earlier slide, I thought NURBS was a better alternative here as it would lower the number of required control points by putting weights on them, without sacrificing the expressibility of the curve..