Commit 054414a5 authored by Guillaume HEUSCH's avatar Guillaume HEUSCH
Browse files

[algorithm] added a minimal SVM version

parent e85eb10b
Pipeline #21035 failed with stage
in 16 minutes and 31 seconds
...@@ -6,6 +6,9 @@ from .SVMCascadePCA import SVMCascadePCA ...@@ -6,6 +6,9 @@ from .SVMCascadePCA import SVMCascadePCA
from .MLP import MLP from .MLP import MLP
from .PadLDA import PadLDA from .PadLDA import PadLDA
from .OCSVM import OCSVM from .OCSVM import OCSVM
from .OCGMM import OCGMM
from .SKLGMM import SKLGMM
from .mySVM import mySVM
def __appropriate__(*args): def __appropriate__(*args):
"""Says object was actually declared here, and not in the import module. """Says object was actually declared here, and not in the import module.
#!/usr/bin/env python
# vim: set fileencoding=utf-8 :
import numpy
from bob.pad.base.algorithm import Algorithm
import bob.learn.libsvm
class mySVM(Algorithm):
This class interfaces a One Class SVM classifier used for PAD
def __init__(self, rescale=True, gamma=0.1, **kwargs):
self.rescale = rescale
self.gamma = gamma
self.machine = None
self.trainer = bob.learn.libsvm.Trainer(machine_type='C_SVC', kernel_type='RBF', probability=True)
setattr(self.trainer, 'gamma', self.gamma)
def train_projector(self, training_features, projector_file):
Trains a SVM
# training_features[0] - training features for the REAL class.
# training_features[1] - training features for the ATTACK class.
# The data - "positive class only"
pos = numpy.array(training_features[0])
neg = numpy.array(training_features[1])
if self.rescale:
for i in range(pos.shape[0]):
min_value = numpy.min(pos[i])
max_value = numpy.max(pos[i])
pos[i] = ((2 * (pos[i] - min_value))/ (max_value - min_value)) - 1
for i in range(neg.shape[0]):
min_value = numpy.min(neg[i])
max_value = numpy.max(neg[i])
neg[i] = ((2 * (neg[i] - min_value))/ (max_value - min_value)) - 1
data = [pos, neg]
# train
self.machine = self.trainer.train(data)
f =, 'w')
def project(self, feature):
Project the given feature
if self.rescale:
min_value = numpy.min(feature)
max_value = numpy.max(feature)
feature = ((2 * (feature - min_value))/ (max_value - min_value)) - 1
return self.machine(feature)
def score(self, toscore):
return [toscore[0]]
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment