ENH Use color histograms with SVMs+Grid Search

This achieves better classification than the Sobel example, while also
being easier to explain.

Author: luispedro

ch10/README.rst

@@ -33,7 +33,7 @@ features.py
     Contains the ``edginess_sobel`` function from the book as well as a simple
     wrapper around ``mahotas.texture.haralick``
 simple_classification.py
-    Classify SimpleImageDataset with texture features + sobel feature
+    Classify SimpleImageDataset with texture features + color histogram features
 figure18.py
     Classify ``AnimTransDistr`` with both texture and SURF features.
 

ch10/simple_classification.py

@@ -7,17 +7,23 @@
 
 import mahotas as mh
 from sklearn import cross_validation
-from sklearn.linear_model.logistic import LogisticRegression
+from sklearn.pipeline import Pipeline
+from sklearn.preprocessing import StandardScaler
+from sklearn.svm import SVC
+from sklearn.grid_search import GridSearchCV
+
+
 import numpy as np
 from glob import glob
-from features import texture, edginess_sobel
+from features import texture, edginess_sobel, color_histogram
 
 basedir = '../SimpleImageDataset/'
 
 
 haralicks = []
 sobels = []
 labels = []
+chists = []
 
 print('This script will test (with cross-validation) classification of the simple 3 class dataset')
 print('Computing features...')
@@ -29,8 +35,10 @@
 # ordering that the filesystem uses
 for fname in sorted(images):
     im = mh.imread(fname, as_grey=True)
+    imc = mh.imread(fname)
     haralicks.append(texture(im))
     sobels.append(edginess_sobel(im))
+    chists.append(color_histogram(imc))
 
     # Files are named like building00.jpg, scene23.jpg...
     labels.append(fname[:-len('xx.jpg')])
@@ -40,25 +48,56 @@
 haralicks = np.array(haralicks)
 sobels = np.array(sobels)
 labels = np.array(labels)
+chists = np.array(chists)
+
+haralick_plus_sobel = np.hstack([np.atleast_2d(sobels).T, haralicks])
+haralick_plus_chists = np.hstack([chists, haralicks])
+haralick_plus_chists_plus_sobel = np.hstack([chists, haralicks, np.atleast_2d(sobels).T])
 
-# We use logistic regression because it is very fast.
+cv=cross_validation.LeaveOneOut(len(images))
+
+# We use SVM because it achieves high accuracy on small(ish) datasets
 # Feel free to experiment with other classifiers
+C_range = 10.0 ** np.arange(-2, 9)
+gamma_range = 10.0 ** np.arange(-5, 4)
+param_grid = dict(gamma=gamma_range, C=C_range)
+grid = GridSearchCV(SVC(), param_grid=param_grid)
+
+p = Pipeline([('preproc', StandardScaler()),
+                ('classifier', grid)])
+
+scores = cross_validation.cross_val_score(
+    p, haralicks, labels, cv=cv)
+print('Accuracy (5 fold x-val) with Logistic Regrssion [haralick features]: {}%'.format(
+    0.1 * round(1000 * scores.mean())))
+
+scores = cross_validation.cross_val_score(
+    p, chists, labels, cv=cv)
+print('Accuracy (5 fold x-val) with Logistic Regrssion [color histograms]: {}%'.format(
+    0.1 * round(1000 * scores.mean())))
+
 scores = cross_validation.cross_val_score(
-    LogisticRegression(), haralicks, labels, cv=5)
-print('Accuracy (5 fold x-val) with Logistic Regrssion [std features]: {}%'.format(
+    p, haralick_plus_chists, labels, cv=cv)
+print('Accuracy (5 fold x-val) with Logistic Regrssion [texture features + chists]: {}%'.format(
     0.1 * round(1000 * scores.mean())))
 
 haralick_plus_sobel = np.hstack([np.atleast_2d(sobels).T, haralicks])
 scores = cross_validation.cross_val_score(
-    LogisticRegression(), haralick_plus_sobel, labels, cv=5).mean()
-print('Accuracy (5 fold x-val) with Logistic Regrssion [std features + sobel]: {}%'.format(
+    p, haralick_plus_sobel, labels, cv=cv)
+print('Accuracy (5 fold x-val) with Logistic Regrssion [texture features + sobel]: {}%'.format(
+    0.1 * round(1000 * scores.mean())))
+
+haralick_plus_chists_plus_sobel = np.hstack([np.atleast_2d(sobels).T, haralicks])
+scores = cross_validation.cross_val_score(
+    p, haralick_plus_chists_plus_sobel, labels, cv=cv)
+print('Accuracy (5 fold x-val) with Logistic Regrssion [texture features + color histograms + sobel]: {}%'.format(
     0.1 * round(1000 * scores.mean())))
 
 
 # We can try to just use the sobel feature. The result is almost completely
 # random.
 scores = cross_validation.cross_val_score(
-    LogisticRegression(), np.atleast_2d(sobels).T, labels, cv=5).mean()
+    p, np.atleast_2d(sobels).T, labels, cv=cv).mean()
 print('Accuracy (5 fold x-val) with Logistic Regrssion [only using sobel feature]: {}%'.format(
     0.1 * round(1000 * scores.mean())))