Feature Lines Extraction

FeatureLine is used to extract feature lines as per the paper (Gumhold, Stefan & Wang, Xinlong & MacLeod, Rob. (2001). Feature Extraction from Point Clouds. Proceedings of 10th international meshing roundtable. 2001. )

install requirements

after creating the environment

foo@bar:~$ conda install --file requirements.txt

or

foo@bar:~$ pip3 install requirements.txt

Usage

from helper import FeatureLines

# returns 'Object after the downsampling specified'
Obj1 = FeatureLines("data/group19/fragment.ply",voxel_size=0.1)

# calculates all penalty functions after creating the graph where N = 16 (nearest neighbor)
Obj1.init(16)

Penalties calculated

w_k : curvature estimate for each node (one value per node)

w_cr_v : vector valued penalty on how well the eigenvalues fit to the crease ( 3D vector value per node )

w_b1 : vector valued border penalty function ( 3D vector value per node )

w_b2 : probability of being border point ( one value per node )

w_co : The corner penalty function ( one value per node )

Show Results

data = Obj1.w_k
Obj1.show_heat(data)

data = Obj1.w_cr_v 
Obj1.show_heat(data)

data = Obj1.w_b1
Obj1.show_heat(data)

data = Obj1.w_b2
Obj1.show_heat(data)

data = Obj1.w_co
Obj1.show_heat(data)

Experiment module

Default Defined patterns Creation

from runner import experiment

## load the config file for the expirments 
my_expirments = experiment("conf1.yaml")

## run all expirments in parallel depending on the number of cores avaliable 
my_expirments.run()
## show results
my_expirments.results