metallic_classifier.ipynb
{
"metadata": {
"name": ""
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "code",
"collapsed": false,
"input": [
"import glob, os, sys\n",
"import operator\n",
"import glob\n",
"\n",
"from skimage import color as col\n",
"\n",
"#from colormath.color_objects import RGBColor\n",
"from PIL import Image\n",
"from StringIO import StringIO\n",
"import numpy as np\n",
"import pylab as pl\n",
"from sklearn import svm, metrics, datasets"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 13
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def histeq(im, nbr_bins=256):\n",
" #get image histogram\n",
" imhist,bins = histogram(im.flatten(),nbr_bins,density=False)\n",
" cdf = imhist.cumsum() #cumulative distribution function\n",
" cdf = 255 * cdf / cdf[-1] #normalize\n",
"\n",
" #print bins.shape\n",
" #use linear interpolation of cdf to find new pixel values\n",
" im2 = interp(im.flatten(),bins[:-1],cdf)\n",
"\n",
" return im2.reshape(im.shape)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 90
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def get_ms(rgb):\n",
" m = (1.0/3.0) * float((sum(rgb)))\n",
" b_r = rgb[0] + rgb[2]\n",
"\n",
" if(b_r >= 2 * rgb[1]):\n",
" s = (3.0 / 2.0) * (rgb[0] - m)\n",
" else:\n",
" s = (3.0 / 2.0) * (m - rgb[2])\n",
" return [int(m),int(s)]"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 72
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def transform_ms(img):\n",
" PIXEL_SAMPLING_INTERVAL = 1\n",
" w, h, r = img.shape\n",
" transformed = np.zeros( (w,h,2), dtype=np.uint8)\n",
" for x in xrange(0, w, PIXEL_SAMPLING_INTERVAL):\n",
" for y in xrange(0, h, PIXEL_SAMPLING_INTERVAL):\n",
" transformed[x,y] = get_ms(img[x,y])\n",
"\n",
" transformed[:,:,1] = histeq(transformed[:,:,1])\n",
" return transformed"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 92
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def get_img_data(img):\n",
" subset_len = 5\n",
" #img_subset = transform_ms(img[:subset_len,:subset_len])\n",
" img_subset = col.rgb2lab(img[:subset_len,:subset_len])\n",
" #img_subset = img[:subset_len,:subset_len]\n",
" #print img_subset.shape\n",
" img_subset = img_subset.reshape(-1)\n",
" #print img_subset.shape\n",
" return img_subset"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 99
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"metals_dir = \"metal_shoes_swatch\"\n",
"non_metals_dir = \"non_metal_shoes_swatch\"\n",
"\n",
"metal_imgs = [get_img_data(pl.imread(img)) for img in glob.glob(metals_dir + \"/*.jpg\")]\n",
"metal_targets = [\"metal\" for i in range(0,len(metal_imgs))]\n",
"#print(metal_targets[0])\n",
"#print len(metal_imgs)\n",
"non_metal_imgs = [get_img_data(pl.imread(img)) for img in glob.glob(non_metals_dir + \"/*.jpg\")]\n",
"non_metal_targets = [\"non\" for i in range(0,len(non_metal_imgs))]\n",
"#print len(non_metal_imgs)\n",
"#print metal_imgs[0].shape\n",
"\n",
"#metal_data = [get_data(img) for img in metal_imgs]\n",
"\n",
"train_num = 50\n",
"\n",
"train_data = metal_imgs[:train_num]\n",
"test_data = metal_imgs[train_num:]\n",
"train_data = train_data + non_metal_imgs[:train_num]\n",
"test_data = test_data + non_metal_imgs[train_num:]\n",
"\n",
"train_targets = metal_targets[:train_num]\n",
"test_targets = metal_targets[train_num:]\n",
"\n",
"train_targets = train_targets + non_metal_targets[:train_num]\n",
"test_targets = test_targets + non_metal_targets[train_num:]\n",
"\n",
"print train_targets\n",
"\n",
"\n",
"digits = datasets.load_digits()\n",
"\n",
"# To apply an classifier on this data, we need to flatten the image, to\n",
"# turn the data in a (samples, feature) matrix:\n",
"#n_samples = len(digits.images)\n",
"#print digits.images.shape\n",
"#data = digits.images.reshape((n_samples, -1))\n",
"#print(n_samples)\n",
"#print data[0].shape"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"['metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'metal', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non', 'non']\n"
]
}
],
"prompt_number": 100
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"classifier = svm.SVC(gamma=0.001)\n",
"# We learn the digits on the first half of the digits\n",
"#classifier.fit(data[:n_samples / 2], digits.target[:n_samples / 2])\n",
"classifier.fit(train_data, train_targets)\n",
"\n",
"# Now predict the value of the digit on the second half:\n",
"expected = test_targets\n",
"predicted = classifier.predict(test_data)\n",
"print predicted\n",
"\n",
"print(\"Classification report for classifier %s:\\n%s\\n\"\n",
" % (classifier, metrics.classification_report(expected, predicted)))\n",
"print(\"Confusion matrix:\\n%s\" % metrics.confusion_matrix(expected, predicted))\n"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"['metal' 'metal' 'non' 'non' 'non' 'non' 'non' 'non' 'non' 'metal' 'metal'\n",
" 'non' 'non' 'non' 'non' 'metal' 'metal' 'metal' 'non' 'non' 'non' 'non'\n",
" 'metal' 'non' 'non' 'non' 'non' 'metal' 'non' 'non' 'non' 'non' 'non'\n",
" 'non' 'non' 'non' 'non' 'non' 'non' 'non']\n",
"Classification report for classifier SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0, degree=3,\n",
" gamma=0.001, kernel=rbf, max_iter=-1, probability=False,\n",
" random_state=None, shrinking=True, tol=0.001, verbose=False):\n",
" precision recall f1-score support\n",
"\n",
" metal 0.78 0.35 0.48 20\n",
" non 0.58 0.90 0.71 20\n",
"\n",
"avg / total 0.68 0.62 0.59 40\n",
"\n",
"\n",
"Confusion matrix:\n",
"[[ 7 13]\n",
" [ 2 18]]\n"
]
}
],
"prompt_number": 101
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
}
],
"metadata": {}
}
]
}