diff --git a/README.rst b/README.rst
index 990a422a35416f574a52eeb4d8f91aa8d4fb8bce..c781c327bcc69802c6dfab78c0debfd5676d3b2d 100644
--- a/README.rst
+++ b/README.rst
@@ -13,6 +13,9 @@ have a Bob version yet, you can get it from:
 
   http://idiap.github.com/bob/
   
+If you already have installed Bob, please make sure that you have at least 
+the version 1.0.3, otherwise the example won't work.
+  
 The AT&T image database is quite small, but sufficient to show how the face
 verification methods work. Still, the results may not be meaningful. One good
 thing about the AT&T database is that it is freely available. You can download
diff --git a/faceverify/eigenface.py b/faceverify/eigenface.py
index 753b93bc94d6f67094eef5242e8d52972f64f5f8..01d20cd7701ed002128c160d3e8e4e1af66dbffd 100644
--- a/faceverify/eigenface.py
+++ b/faceverify/eigenface.py
@@ -85,6 +85,15 @@ def main():
   print "Training PCA machine"
   pca_machine = train(training_images)
   
+  # extract the first eigenvector
+  projection_matrix = pca_machine.weights
+  eigen_image = projection_matrix[:,0]
+  # reshape the eigenvector to the right image resolution
+  eigen_image.shape = (112,92)
+  # show the first eigenface
+  pyplot.imshow(eigen_image, cmap = 'gray')
+  pyplot.show()
+  
   #####################################################################
   ### extract eigenface features of model and probe images
 
diff --git a/localbob.cfg b/localbob.cfg
index f51b78b83c83fd46371a5abc1c759363b7e635e7..acdef6fa45afc4d636f743fb071f92dc23ad6b6c 100644
--- a/localbob.cfg
+++ b/localbob.cfg
@@ -36,7 +36,7 @@ recipe = local.bob.recipe:config
 ; If you go for a relative path, remember this path is relative to the path
 ; you are calling buildout from. This is normally the path that contains the
 ; file 'setup.py' for your own package.
-install-directory = ../bob/build
+install-directory = /scratch/mguenther/Bob/cmake/trunk
 
 ; Some examples for Idiap:
 ;install-directory = /idiap/group/torch5spro/releases/bob-1.0.2/install/linux-x86_64-release