diff --git a/bob/bio/vein/configurations/principal_curvature.py b/bob/bio/vein/configurations/principal_curvature.py
new file mode 100644
index 0000000000000000000000000000000000000000..8357ea2a5a92e104a63f7df2312abe7413704377
--- /dev/null
+++ b/bob/bio/vein/configurations/principal_curvature.py
@@ -0,0 +1,50 @@
+#!/usr/bin/env python
+# vim: set fileencoding=utf-8 :
+# Tue 03 Nov 2020 16:48:32 CEST
+
+'''Principal Curvature and Miura Matching baseline
+
+References:
+
+1. [CW09]_
+2. [TV13]_
+3. [TVM14]_
+
+'''
+
+sub_directory = 'pc'
+"""Sub-directory where results will be placed.
+
+You may change this setting using the ``--sub-directory`` command-line option
+or the attribute ``sub_directory`` in a configuration file loaded **after**
+this resource.
+"""
+
+from ..preprocessor import NoCrop, TomesLeeMask, HuangNormalization, \
+ NoFilter, Preprocessor
+
+preprocessor = Preprocessor(
+ crop=NoCrop(),
+ mask=TomesLeeMask(),
+ normalize=HuangNormalization(),
+ filter=NoFilter(),
+ )
+"""Preprocessing using gray-level based finger cropping and no post-processing
+"""
+
+from ..extractor import PrincipalCurvature
+extractor = PrincipalCurvature()
+"""Features are the output of the principal curvature algorithm, as described on
+[CW09]_.
+
+Defaults taken from [CW09]_.
+"""
+
+# Notice the values of ch and cw are different than those from the
+# repeated-line tracking baseline.
+from ..algorithm import MiuraMatch
+algorithm = MiuraMatch()
+"""Miura-matching algorithm with specific settings for search displacement
+
+Defaults taken from [TV13]_.
+"""
diff --git a/bob/bio/vein/extractor/PrincipalCurvature.py b/bob/bio/vein/extractor/PrincipalCurvature.py
index c08cae315ea7e2c571cfe17565372f50962cc24c..3a8f97f324d7b51e655b0fcf7799ef1a77604385 100644
--- a/bob/bio/vein/extractor/PrincipalCurvature.py
+++ b/bob/bio/vein/extractor/PrincipalCurvature.py
@@ -7,87 +7,97 @@ import bob.io.base
from bob.bio.base.extractor import Extractor
+from scipy.ndimage import gaussian_filter
+
class PrincipalCurvature (Extractor):
- """MiuraMax feature extractor
+ """MiuraMax feature extractor
- Based on J.H. Choi, W. Song, T. Kim, S.R. Lee and H.C. Kim, Finger vein
- extraction using gradient normalization and principal curvature. Proceedings
- on Image Processing: Machine Vision Applications II, SPIE 7251, (2009)
- """
+ Based on J.H. Choi, W. Song, T. Kim, S.R. Lee and H.C. Kim, Finger vein
+ extraction using gradient normalization and principal curvature. Proceedings
+ on Image Processing: Machine Vision Applications II, SPIE 7251, (2009)
+ """
- def __init__(
- self,
- sigma = 2, # Gaussian standard deviation applied
- threshold = 1.3, # Percentage of maximum used for hard thresholding
- ):
+ def __init__(
+ self,
+ sigma = 3, # Gaussian standard deviation applied
+ threshold = 2, # Percentage of maximum used for hard thresholding
+ ):
+ """NOTE: In the reference paper where the size of the finger image is 320 by 128,
+ the proposed values for sigma and threshold are 3 and 4, respectively.
+ However, for other resolutions it is better to change the values for sigma and
+ threshold. e.g., in UTFVP dataset where the size of the finger image is 672 by 380,
+ sigma=6 and threshold=4 workes better results better features.
+ """
+ # call base class constructor
+ Extractor.__init__(
+ self,
+ sigma = sigma,
+ threshold = threshold,
+ )
- # call base class constructor
- Extractor.__init__(
- self,
- sigma = sigma,
- threshold = threshold,
- )
+ # block parameters
+ self.sigma = sigma
+ self.threshold = threshold
- # block parameters
- self.sigma = sigma
- self.threshold = threshold
+ def ut_gauss(self, img, sigma, dx, dy):
+ return gaussian_filter(numpy.float64(img), sigma, order = [dx,dy])
- def principal_curvature(self, image, mask):
- """Computes and returns the Maximum Curvature features for the given input
- fingervein image"""
+ def principal_curvature(self, image, mask):
+ """Computes and returns the Maximum Curvature features for the given input
+ fingervein image"""
- finger_mask = numpy.zeros(mask.shape)
- finger_mask[mask == True] = 1
+ finger_mask = numpy.zeros(mask.shape)
+ finger_mask[mask == True] = 1
- sigma = numpy.sqrt(self.sigma**2/2)
+ sigma = numpy.sqrt(self.sigma**2/2)
- gx = ut_gauss(img,sigma,1,0)
- gy = ut_gauss(img,sigma,0,1)
+ gx = self.ut_gauss(image,self.sigma,1,0)
+ gy = self.ut_gauss(image,self.sigma,0,1)
- Gmag = numpy.sqrt(gx**2 + gy**2) # Gradient magnitude
+ Gmag = numpy.sqrt(gx**2 + gy**2) # Gradient magnitude
- # Apply threshold
- gamma = (self.threshold/100)*max(max(Gmag))
+ # Apply threshold
+ gamma = (self.threshold/100)*numpy.max(Gmag)
- indices = find(Gmag < gamma)
+ indices = numpy.where(Gmag < gamma)
- gx[indices] = 0
- gy[indices] = 0
+ gx[indices] = 0
+ gy[indices] = 0
- # Normalize
- Gmag[find[Gmag==0]] = 1 # Avoid dividing by zero
- gx = gx/Gmag
- gy = gy/Gmag
+ # Normalize
+ Gmag[numpy.where(Gmag==0)] = 1 # Avoid dividing by zero
+ gx = gx/Gmag
+ gy = gy/Gmag
- hxx = ut_gauss(gx,sigma,1,0)
- hxy = ut_gauss(gx,sigma,0,1)
- hyy = ut_gauss(gy,sigma,0,1)
+ hxx = self.ut_gauss(gx,sigma,1,0)
+ hxy = self.ut_gauss(gx,sigma,0,1)
+ hyy = self.ut_gauss(gy,sigma,0,1)
- lambda1 = 0.5*(hxx + hyy + numpy.sqrt(hxx**2 + hyy**2 - 2*hxx**hyy + 4*hxy**2))
- veins = lambda1*finger_mask
- # Normalise
- veins = veins - min(veins[:])
- veins = veins/max(veins[:])
+ lambda1 = 0.5*(hxx + hyy + numpy.sqrt(hxx**2 + hyy**2 - 2*hxx*hyy + 4*hxy**2))
+ veins = lambda1*finger_mask
- veins = veins*finger_mask
+ # Normalise
+ veins = veins - numpy.min(veins[:])
+ veins = veins/numpy.max(veins[:])
+ veins = veins*finger_mask
- # Binarise the vein image by otsu
- md = numpy.median(img_veins[img_veins>0])
- img_veins_bin = img_veins > md
+ # Binarise the vein image by otsu
+ md = numpy.median(veins[veins>0])
+ img_veins_bin = veins > md
- return img_veins_bin.astype(numpy.float64)
+ return img_veins_bin.astype(numpy.float64)
- def __call__(self, image):
- """Reads the input image, extract the features based on Principal Curvature
- of the fingervein image, and writes the resulting template"""
+ def __call__(self, image):
+ """Reads the input image, extract the features based on Principal Curvature
+ of the fingervein image, and writes the resulting template"""
- finger_image = image[0] #Normalized image with or without histogram equalization
- finger_mask = image[1]
+ finger_image = image[0] #Normalized image with or without histogram equalization
+ finger_mask = image[1]
- return self.principal_curvature(finger_image, finger_mask)
+ return self.principal_curvature(finger_image, finger_mask)
\ No newline at end of file
diff --git a/doc/references.rst b/doc/references.rst
index 5265c8692e7e35d4edf62989ca46c4ca551abed2..8e5e16c8d0c672374f0ca175c0b9f368e1cfde34 100644
--- a/doc/references.rst
+++ b/doc/references.rst
@@ -27,3 +27,4 @@
.. [KK10] *R. Kabacinski and *M. Kowalski, **Human Vein Pattern Segmentation from Low Quality Images - A Comparison of Methods**, in: Image Processing and Communications Challenges 2, pp. 105-112, 2010.
+.. [CW09] J.H. Choi, W. Song, T. Kim, S.R. Lee and H.C. Kim, **Finger vein extraction using gradient normalization and principal curvature**, Proceedings on Image Processing: Machine Vision Applications II, SPIE 7251, 2009.
\ No newline at end of file
diff --git a/setup.py b/setup.py
index 986e76be7263be8ea30a5e749641668d0e388d0c..f6427c731e7a7305e1b63ec1609322b585c92500 100644
--- a/setup.py
+++ b/setup.py
@@ -40,6 +40,7 @@ setup(
# baselines
'mc = bob.bio.vein.configurations.maximum_curvature',
+ 'pc = bob.bio.vein.configurations.principal_curvature',
'rlt = bob.bio.vein.configurations.repeated_line_tracking',
'wld = bob.bio.vein.configurations.wide_line_detector',