Can we assume that no two triangles are co-planar? about computer-graphics-bounding-volume-hierarchy

[zhtdavis]

How long does it take for you to do brute force intersection test? Mine is around 7s. On the readme it seems the demo program is using the second algorithm which is much faster.

[zhtdavis]

Mine is like:

Method	Time in seconds
brute force	7.68331193924
build trees	0.01909399033
use trees	0.00568318367

[velociraptor111]

yeah, mine takes a long time as well like ~11 seconds.
Did you use algorithm 1 as well?

[zhtdavis]

Yes.

[velociraptor111]

Just to verify if I'm doing this correctly, you check for these conditions:

1. If two edges from triangle A intersect triangle B && 0 edges from triangle B intersect triangle A return true
2. If two edges from triangle B intersect triangle A && 2 edges from triangle A intersect triangle B
   return true
3. If 1 edge from triangle A intersect triangle B && 1 edge from triangle B intersect triangle A
   return true

else return false

is that all?

[zhtdavis]

I'm doing six ray-triangle intersection queries. Not sure this is the correct way to do it...

[velociraptor111]

Yeah same, so I converted each edge basically into a ray and use the function from "ray_intersect_triangle.h" to intersect the with "ray" with the triangle. Ray origin if one of the vertex (call this A) and the direction is going from this ray to the next vertex B. (hence direction is B-A)

If ray intersects the triangle then it means that edge intersects the triangle.

not sure either lol

[krsh732]

So do we need to worry about coplanar triangles?

[velociraptor111]

I got the correct results .. So I think yes ..(?) also the chances of triangles exactly being coplanar are pretty slim ...