lena = lena() / 256.0 height, width = lena.shape # Extract all clean patches from the left half of the image print 'Extracting clean patches...' t0 = time() patch_size = (6, 6) data = extract_patches_2d(lena, patch_size) data = data.reshape(data.shape[0], -1) data -= np.mean(data, axis=0) data /= np.std(data, axis=0) print 'done in %.2fs.' % (time() - t0) ############################################################################### # Learn the dictionary from clean patches print 'Learning the dictionary SparseCoding... ' t0 = time() np.random.seed(42) dico = SparseCoding(n_atoms=36, alpha=.1, max_iter=200) V = dico.fit(data).components_ print 'done in %.2fs.' % (time() - t0) print 'Transform the data...' t0 = time() coef = dico.transform(data[:10,:]) print 'done in %.2fs.' % (time() - t0) np.testing.assert_almost_equal(coef[:10,:].sum(),3.23775956895) np.testing.assert_almost_equal(V[:,:10].sum(),39.9447109218)
# Extract all clean patches from the left half of the image
print 'Extracting clean patches...'
t0 = time()
patch_size = (7, 7)
data = extract_patches_2d(distorted[:, :height / 2], patch_size)
data = data.reshape(data.shape[0], -1)
data -= np.mean(data, axis=0)
data /= np.std(data, axis=0)
print 'done in %.2fs.' % (time() - t0)
###############################################################################
# Learn the dictionary from clean patches (new way)
print 'Learning the dictionary SparseCoding... '
t0 = time()
dico = SparseCoding(n_atoms=100, alpha=1., max_iter=10000)
V = dico.fit(data).components_
dt = time() - t0
print 'done in %.2fs.' % dt
pl.figure(figsize=(4.2, 4))
for i, comp in enumerate(V[:100]):
pl.subplot(10, 10, i + 1)
pl.imshow(comp.reshape(patch_size), cmap=pl.cm.gray_r,
interpolation='nearest')
pl.xticks(())
pl.yticks(())
pl.suptitle('Dictionary learned from Lena patches\n' +
'Train time %.1fs on %d patches' % (dt, len(data)),
fontsize=16)
pl.subplots_adjust(0.08, 0.02, 0.92, 0.85, 0.08, 0.23)
