I am a third year undergraduate student at Berkeley, and I have recently been working on implementing IGM-based quad meshing in Blender using Python. IGM-based methods are particularly interesting to me because from my reading they seem to be very amenable to user control; as a modeler, I am interested in possibilities to guide quad topology, such as maybe using the grease pencil to define contours on the mesh surface. Seeing as GSoC proposals are soon to open, I was wondering if this idea seemed interesting and viable to you all for a GSoC project, even though there are many alternatives out there for quad meshing in Blender already.
A good quad remesher natively integrated in Blender would be very welcome.
For a GSoC project, the most effective thing would be to use an existing implementation and integrate and test it really well, and maybe improve the implementation where needed. These algorithms take a long time to implement in a robust way, which does not leave enough time for good integration.
QuadriFlow is very good from what I can tell. As far as I know all similar algorithm use some kind of orientation field that could be controlled by user. Good performance on high poly meshes is important, sculpting generates models with millions of polygons and some of the published methods are quite slow which is not ideal for feedback on user controlled orientation.
