I had been wondering what edge flip was for when I was implementing it so it's cool to see it in context -- what other uses does edge flip have as an operation?
VegitableChicken
Why does the flip edge result look so different from the one we made?
rmvenkat
I think flip edge seems fine. The Splitting operation seems confusing, because I think the order in which the splitting happens as well as the edges chosen matter a lot. I actually think this version of split is a bit different from what we did.
ryanmelon
For this algorithm, the number of faces will increase by a integer times, which is kind of difficult to control how detailed I want. If there any improvement of this algorithm on solving this issue?
Max
Both the split and flip operations are the same as what you've been working on - you can actually implement loop subdivision in part 2 with them.
Here, the split is applied to each edge of the original mesh in an arbitrary order. It's a little hard to tell, but you can find an ordering of splits that gives the picture. Also note that each edge split creates one or two blue edges depending on the boundary. The flip is also the same - it's only applied to edges between a blue and a black vertex here, which should be a bit easier to see.
@ryanmelon I don't think there is an adaptation to loop subdivision itself, but you could always run a simplification algorithm afterward to remove faces.
Parker
Cool to see we can do something powerful by just putting together some simple stuff.
WonderWoman
why do we get the edges that connect an old and a new vertex from edge split?
rilakkuma
The algorithm seems complicated, but after implementing it, it was quite easy using the halfedge data structure that we have.
aki
It's important to note that in meshEdit, because of the way edges are stored, one of the "new edges" made during edgeSplit is not considered to be new in this description of the algorithm.
Lockbrains
This does show us the function of edge flip, and I didn't it is used for this kind of queries. But that does feels a little bit strange to me because it seems to me that not all the vertices are considered in this flipping? Some of the vertices are not marked by blue circles in the picture, and it seems to me they are not even considered in calculation. Why?
I had been wondering what edge flip was for when I was implementing it so it's cool to see it in context -- what other uses does edge flip have as an operation?
Why does the flip edge result look so different from the one we made?
I think flip edge seems fine. The Splitting operation seems confusing, because I think the order in which the splitting happens as well as the edges chosen matter a lot. I actually think this version of split is a bit different from what we did.
For this algorithm, the number of faces will increase by a integer times, which is kind of difficult to control how detailed I want. If there any improvement of this algorithm on solving this issue?
Both the split and flip operations are the same as what you've been working on - you can actually implement loop subdivision in part 2 with them.
Here, the split is applied to each edge of the original mesh in an arbitrary order. It's a little hard to tell, but you can find an ordering of splits that gives the picture. Also note that each edge split creates one or two blue edges depending on the boundary. The flip is also the same - it's only applied to edges between a blue and a black vertex here, which should be a bit easier to see.
@ryanmelon I don't think there is an adaptation to loop subdivision itself, but you could always run a simplification algorithm afterward to remove faces.
Cool to see we can do something powerful by just putting together some simple stuff.
why do we get the edges that connect an old and a new vertex from edge split?
The algorithm seems complicated, but after implementing it, it was quite easy using the halfedge data structure that we have.
It's important to note that in meshEdit, because of the way edges are stored, one of the "new edges" made during edgeSplit is not considered to be new in this description of the algorithm.
This does show us the function of edge flip, and I didn't it is used for this kind of queries. But that does feels a little bit strange to me because it seems to me that not all the vertices are considered in this flipping? Some of the vertices are not marked by blue circles in the picture, and it seems to me they are not even considered in calculation. Why?