Cutaway visualisations

[Robadob]

Useful for things like tumours where there are alot of densely packed cells.

e.g. https://youtu.be/nJ2urSm4ilU

Done via matlab cutaway plot: https://github.com/MathCancer/PhysiCell/blob/81e85ce1dcb7156864a4460666e60878b82d3e97/matlab/simple_cutaway_plot.m

[MILeach]

Ray casting probably a good option for this - easily integrates CSG operations, intersection, difference etc.

[MILeach]

Ray casting probably a good option for this - easily integrates CSG operations, intersection, difference etc.

If you are looking to have individual spheres sliced as in the example image. Obviously if that isn't required you could just cull them

[Robadob]

We don't currently render back faces, so a vertex dump for vertices/agents in a certain area would look weird, as you could see through the insides of cells. (aka culling) I was considering something similar, where agents on the boundary have any vertices within the banned area are flattened into the boundary. But this wouldn't work nice if say the majority of a sphere was in the banned area. So it does need some thought. Culling agents outside of the area could work, but the 'area' would look rougher. I think the cut through thing would look nice. Possible that something could be done with stencil buffer, but less sure on that.

…

On Mon, 17 May 2021 at 21:42, MILeach ***@***.***> wrote: Ray casting probably a good option for this - easily integrates CSG operations, intersection, difference etc. If you are looking to have individual spheres sliced as in the example image. Obviously if that isn't required you could just cull them — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#62 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFVGCR4IIJB4STGCDNQ7M3TOF5R3ANCNFSM45ARPHQA> .

[Robadob]

I would need to read up more on ray-casting, to consider that (e.g. is perf going to be trash).
This isn't a priority though, just a thought of something that could add value. I'm unlikely to really dig into it further until/if Primage model is fully moved over to fgpu2. I might also pick back up the surface renderer I was building then too.

[MILeach]

Oops meant raymarching - this video gives a nice demo and explanation

https://youtu.be/Cp5WWtMoeKg

[Robadob]

Watching that, sounds as though ray marching would do the job. But it would also be a major rewrite of the visualiser, due to the processing required (e.g. doesn't seem suitable for basic multi-pass rendering with potentially millions of agents).