Implemented SVM classifier and added test file to test it.

.gitignore

@@ -7,3 +7,14 @@
 *.synctex.gz
 *.toc
 *.out
+*.jpg
+images/BBB
+images/Images
+images/Infos
+images/licenseplates
+chars
+learning_set
+test_set
+classifier
+classifier-model
+classifier-characters

src/Character.py

-# TODO cleanup the getElements stuff
+from LocalBinaryPatternizer import LocalBinaryPatternizer
+
 class Character:
     def __init__(self, value, corners, image):
         self.value   = value
         self.corners = corners
         self.image   = image
-        
+
     # Testing purposes
     def show(self):
         from pylab import imshow, show
@@ -12,4 +13,6 @@ class Character:
         show()
 
     def get_feature_vector(self):
-        pass
+        pattern = LocalBinaryPatternizer(self.image)
+
+        return pattern.create_features_vector()

src/Classifier.py

-from svmutil import svm_model, svm_problem, svm_parameter, svm_predict, LINEAR
-from cPicle import dump, load
+from svmutil import svm_train, svm_problem, svm_parameter, svm_predict, \
+        LINEAR, svm_save_model, svm_load_model
+from cPickle import dump, load
 
 
 class Classifier:
     def __init__(self, c=None, filename=None):
         if filename:
             # If a filename is given, load a modl from the fiven filename
-            f = file(filename, 'r')
-            self.model, self.param, self.character_map = load(f)
+            self.model = svm_load_model(filename + '-model')
+            f = file(filename + '-characters', 'r')
+            self.character_map = load(f)
             f.close()
         else:
             self.param = svm_parameter()
@@ -18,8 +20,9 @@ class Classifier:
 
     def save(self, filename):
         """Save the SVM model in the given filename."""
-        f = file(filename, 'w+')
-        dump((self.model, self.param, self.character_map), f)
+        svm_save_model(filename + '-model', self.model)
+        f = file(filename + '-characters', 'w+')
+        dump(self.character_map, f)
         f.close()
 
     def train(self, learning_set):
@@ -27,8 +30,11 @@ class Classifier:
         known values."""
         classes = []
         features = []
+        l = len(learning_set)
 
-        for char in learning_set:
+        for i, char in enumerate(learning_set):
+            print 'Training "%s"  --  %d of %d (%d%% done)' \
+                    % (char.value, i + 1, l, int(100 * (i + 1) / l))
             # Map the character to an integer for use in the SVM model
             if char.value not in self.character_map:
                 self.character_map[char.value] = len(self.character_map)
@@ -36,15 +42,13 @@ class Classifier:
             classes.append(self.character_map[char.value])
             features.append(char.get_feature_vector())
 
-        problem = svm_problem(self.c, features)
-        self.model = svm_model(problem, self.param)
-
-        # Add prediction function that returns a numeric class prediction
-        self.model.predict = lambda self, x: svm_predict([0], [x], self)[0][0]
+        problem = svm_problem(classes, features)
+        self.model = svm_train(problem, self.param)
 
     def classify(self, character):
         """Classify a character object and assign its value."""
-        prediction = self.model.predict(character.get_feature_vector())
+        predict = lambda x: svm_predict([0], [x], self.model)[0][0]
+        prediction = predict(character.get_feature_vector())
 
         for value, svm_class in self.character_map.iteritems():
             if svm_class == prediction:

src/ClassifierTest.py

+#!/usr/bin/python
+from LicensePlate import LicensePlate
+from Classifier import Classifier
+from cPickle import dump, load
+
+#chars = []
+#
+#for i in range(9):
+#    for j in range(100):
+#        try:
+#            filename = '%04d/00991_%04d%02d.info' % (i, i, j)
+#            print 'loading file "%s"' % filename
+#            plate = LicensePlate(i, j)
+#
+#            if hasattr(plate, 'characters'):
+#                chars.extend(plate.characters)
+#        except:
+#            print 'epic fail'
+#
+#print 'loaded %d chars' % len(chars)
+#
+#dump(chars, file('chars', 'w+'))
+#----------------------------------------------------------------
+#chars = load(file('chars', 'r'))
+#learned = []
+#learning_set = []
+#test_set = []
+#
+#for char in chars:
+#    if learned.count(char.value) > 80:
+#        test_set.append(char)
+#    else:
+#        learning_set.append(char)
+#        learned.append(char.value)
+#
+#dump(learning_set, file('learning_set', 'w+'))
+#dump(test_set, file('test_set', 'w+'))
+#----------------------------------------------------------------
+learning_set = load(file('learning_set', 'r'))
+
+# Train the classifier with the learning set
+classifier = Classifier(c=3)
+classifier.train(learning_set)
+#classifier.save('classifier')
+#----------------------------------------------------------------
+#classifier = Classifier(filename='classifier')
+#test_set = load(file('test_set', 'r'))
+#l = len(test_set)
+#matches = 0
+#
+#for i, char in enumerate(test_set):
+#    prediction = classifier.classify(char)
+#
+#    if char.value == prediction:
+#        print ':) ------> Successfully recognized "%s"' % char.value
+#        matches += 1
+#    else:
+#        print ':( Expected character "%s", got "%s"' \
+#                % (char.value, prediction),
+#
+#    print '  --  %d of %d (%d%% done)' % (i + 1, l, int(100 * (i + 1) / l))
+#
+#print '\n%d matches (%d%%), %d fails' % (matches, \
+#        int(100 * matches / len(test_set)), \
+#        len(test_set) - matches)

src/LicensePlate.py

 from pylab import array, zeros, inv, dot, svd, shape, floor
-from xml.dom.minidom import parse 
+from xml.dom.minidom import parse
 from Error import Error
 from Point import Point
 from Character import Character
 from GrayscaleImage import GrayscaleImage
+from NormalizedCharacterImage import NormalizedCharacterImage
 
 '''
     Creates a license plate object based on an XML file. The image should be
     placed in a folder 'images' the xml file in a folder 'xml'
-    
-    TODO: perhaps remove non required XML lookups 
+
+    TODO: perhaps remove non required XML lookups
 '''
 class LicensePlate:
 
-    def __init__(self, xml_title):
-        try:
-            self.dom = parse('../XML/' + str(xml_title))
-        except IOError:
-            Error("Incorrect file name given.")
-        else:
-            properties = self.get_properties()
+    def __init__(self, folder_nr, file_nr):
+        filename = '%04d/00991_%04d%02d' % (folder_nr, folder_nr, file_nr)
+
+        self.dom = parse('../images/Infos/%s.info' % filename)
+        properties = self.get_properties()
+
+        self.image = GrayscaleImage('../images/Images/%s.jpg' % filename)
+        self.width = int(properties['width'])
+        self.height = int(properties['height'])
 
-            self.image  = GrayscaleImage('../images/' + str(properties['uii']) + '.' + str(properties['type']))
-            self.width  = int(properties['width'])
-            self.height = int(properties['height'])
+        self.read_xml()
 
-            self.read_xml()
-            
     # sets the entire license plate of an image
-    def retrieve_data(self, corners):    
+    def retrieve_data(self, corners):
         x0, y0 = corners[0].to_tuple()
         x1, y1 = corners[1].to_tuple()
         x2, y2 = corners[2].to_tuple()
         x3, y3 = corners[3].to_tuple()
-        
+
         M = max(x0, x1, x2, x3) - min(x0, x1, x2, x3)
         N = max(y0, y1, y2, y3) - min(y0, y1, y2, y3)
-                    
+
         matrix = array([
           [x0, y0, 1,  0,  0, 0,    0,    0, 0],
           [ 0,  0, 0, x0, y0, 1,    0,    0, 0],
@@ -44,10 +43,10 @@ class LicensePlate:
           [ 0,  0, 0, x1, y1, 1,    0,    0, 0],
           [x2, y2, 1,  0,  0, 0, -M*x2, -M*y2, -M],
           [ 0,  0, 0, x2, y2, 1, -N*x2, -N*y2, -N],
-          [x3, y3, 1,  0,  0, 0,    0,    0, 0], 
+          [x3, y3, 1,  0,  0, 0,    0,    0, 0],
           [ 0,  0, 0, x3, y3, 1, -N*x3, -N*y3, -N]
         ])
-        
+
         P = inv(self.get_transformation_matrix(matrix))
         data = array([zeros(M, float)] * N)
 
@@ -56,23 +55,23 @@ class LicensePlate:
                 or_coor   = dot(P, ([[i],[j],[1]]))
                 or_coor_h = or_coor[1][0] / or_coor[2][0], or_coor[0][0] / or_coor[2][0]
                 data[j][i] = self.pV(or_coor_h[0], or_coor_h[1])
-                
+
         return data
-                
+
     def get_transformation_matrix(self, matrix):
-        # Get the vector p and the values that are in there by taking the SVD. 
+        # Get the vector p and the values that are in there by taking the SVD.
         # Since D is diagonal with the eigenvalues sorted from large to small on
-        # the diagonal, the optimal q in min ||Dq|| is q = [[0]..[1]]. Therefore, 
+        # the diagonal, the optimal q in min ||Dq|| is q = [[0]..[1]]. Therefore,
         # p = Vq means p is the last column in V.
         U, D, V = svd(matrix)
         p = V[8][:]
-        
+
         return array([
-            [ p[0], p[1], p[2] ], 
-            [ p[3], p[4], p[5] ], 
+            [ p[0], p[1], p[2] ],
+            [ p[3], p[4], p[5] ],
             [ p[6], p[7], p[8] ]
         ])
-            
+
     def pV(self, x, y):
         image = self.image
 
@@ -85,25 +84,25 @@ class LicensePlate:
             y_low  = floor(y)
             y_high = floor(y + 1)
             x_y    = (x_high - x_low) * (y_high - y_low)
-            
+
             a = x_high - x
             b = y_high - y
             c = x - x_low
             d = y - y_low
-            
+
             return image[x_low,  y_low] / x_y * a * b \
                 + image[x_high,  y_low] / x_y * c * b \
                 + image[x_low , y_high] / x_y * a * d \
                 + image[x_high, y_high] / x_y * c * d
-            
+
         return 0
-    
+
     # Testing purposes
     def show(self):
         from pylab import imshow, show
         imshow(self.data, cmap="gray")
         show()
-        
+
     def get_properties(self):
         children = self.get_children("properties")
 
@@ -120,7 +119,7 @@ class LicensePlate:
     # TODO : create function for location / characters as they do the same
     def read_xml(self):
         children = self.get_children("plate") # most recent version
-        
+
         for child in children:
             if child.nodeName == "regnum":
               self.license_full = child.firstChild.data
@@ -130,7 +129,7 @@ class LicensePlate:
                 self.corners = self.get_corners(child)
             elif child.nodeName == "characters":
                 nodes = child.childNodes
-  
+
                 self.characters = []
 
                 for character in nodes:
@@ -139,11 +138,11 @@ class LicensePlate:
                     corners = self.get_corners(character)
                     data    = self.retrieve_data(corners)
                     image   = NormalizedCharacterImage(data=data)
-                    
+
                     self.characters.append(Character(value, corners, image))
             else:
                 pass
-            
+
     def get_node(self, node, dom=None):
         if not dom:
             dom = self.dom
@@ -152,14 +151,14 @@ class LicensePlate:
 
     def get_children(self, node, dom=None):
         return self.get_node(node, dom).childNodes
-        
+
     def get_corners(self, child):
       nodes = self.get_children("quadrangle", child)
-          
+
       corners = []
 
       for corner in nodes:
         if corner.nodeName == "point":
             corners.append(Point(corner))
-            
+
       return corners

src/LocalBinaryPatternizer.py

@@ -3,7 +3,7 @@ from numpy import zeros, byte
 from math import ceil
 
 class LocalBinaryPatternizer:
-        
+
     def __init__(self, image, cell_size=16):
         self.cell_size = cell_size
         self.image = image
@@ -23,7 +23,7 @@ class LocalBinaryPatternizer:
             at each neighbour starting at 7 in the top-left corner. This gives a
             8-bit feature number of a pixel'''
         for y, x, value in self.image:
-            
+
             pattern = (self.is_pixel_darker(y - 1, x - 1, value) << 7) \
                     | (self.is_pixel_darker(y - 1, x    , value) << 6) \
                     | (self.is_pixel_darker(y - 1, x + 1, value) << 5) \
@@ -32,17 +32,17 @@ class LocalBinaryPatternizer:
                     | (self.is_pixel_darker(y + 1, x    , value) << 2) \
                     | (self.is_pixel_darker(y + 1, x - 1, value) << 1) \
                     | (self.is_pixel_darker(y    , x - 1, value) << 0)
-                    
+
             cy, cx = self.get_cell_index(y, x)
             self.features[cy][cx].add(pattern)
 
         return self.get_features_as_array()
-    
+
     def is_pixel_darker(self, y, x, value):
         return self.image.in_bounds(y, x) and self.image[y, x] > value
-        
+
     def get_cell_index(self, y, x):
         return (y / self.cell_size, x / self.cell_size)
-        
+
     def get_features_as_array(self):
-        return [item for sublist in self.features for item in sublist]
+        return [h.bins for h in [h for sub in self.features for h in sub]][0]
\ No newline at end of file