yaoyx689 / amm_nrr Goto Github PK

Hello,

I tried a little AMM_NRR with the provided data.

fixmin method is quite fast. but seems no modification (deformation) on the src_mesh. Is it correct? Or wrong parameters were used?

Which method is yours?

Thank you!

`
USE Welsch fix min
data use welsc = 1
reg_use welsch = 1
use_AA = 0
use_lbfgs = 0
use_dynamic = 0
uni_sample_radio = 5
filename = data/d2/source.obj
Triangle Mesh.
Information of the input mesh:
Vertex : 7345;
Face : 13397;
Edge : 20741, HalfEdge : 41482

filename = data/d2/target.obj
Triangle Mesh.
Information of the input mesh:
Vertex : 7776;
Face : 14040;
Edge : 21829, HalfEdge : 43658

scale = 2.23967

rigid registration to initial...
rgid registration...
non-rigid registration to initial...
non-rigid registration...
wdata = 2.07751e-05 wreg = 0.000721683 wrot = 0.362151 nu1 = 0.00322297 0.00418676 end nu1 = 0.00322297
Registration done!
rigid_init time : 0.0054216 s rigid-reg run time = 0.0001923 s
non-rigid init time = 0.108981 s non-rigid run time = 0.0260164 s
`

Hi the result is bad without landmarkFile param

using
./AMM_NRR <srcFile> <tarFile> <outPath> <radius> <alpha> <beta>
the obj file is blew
obj.zip

could you help me how to run correctly ?

Hi

Thanks for sharing the code.
I was confused at several places in this code as inconsistent with the paper, until I read another paper QNS.
I think this code is kind hybrid-implementation for these two papers, right?
But why did you that? Would it be more clear to test upon pure implementation for ONE paper to verify the advantages and disadvantages of the methods proposed in that paper?

Regards, ymcl