from myHog import hog
import cPickle as pickle
from sklearn import svm
import matplotlib.image as mpimg
from skimage.color import rgb2gray
from skimage import transform
import numpy
filename = "savedLinearSVM19999.pkl"
with open(filename, 'r') as f:
LinearClf1 = pickle.load(f)
images = []
for i in range(61):
test = mpimg.imread("test/" + "test (" + str(i + 1) + ").jpg")
test = transform.resize(test, numpy.array([64, 64]))
test = rgb2gray(test)
test = hog(test).ravel()
images.append(test)
print "...Load OK!"
f = open('test/lables.txt')
data = f.readlines()
f.close
for i in range(len(data)):
data[i] = int(data[i][3])
data = data[0:61]
p = []
correct = 0
for i in range(len(images)):
p.append(LinearClf1.predict(images[i]))
if p[i] == data[i]:
from myHog import hog
import cPickle as pickle
from sklearn import svm
import matplotlib.image as mpimg
from skimage.color import rgb2gray
from skimage import transform
import numpy
filename = "savedLinearSVM19999.pkl"
with open(filename,'r') as f:
LinearClf1 = pickle.load(f)
images = []
for i in range(61):
test = mpimg.imread("test/"+"test ("+ str(i+1) + ").jpg")
test = transform.resize(test,numpy.array([64,64]))
test = rgb2gray(test)
test = hog(test).ravel()
images.append(test)
print "...Load OK!"
f = open('test/lables.txt')
data = f.readlines()
f.close
for i in range(len(data)):
data[i] = int(data[i][3])
data = data[0:61]
p = []
correct = 0
for i in range(len(images)):
p.append(LinearClf1.predict(images[i]))
if p[i] == data[i]:
imgs.append(ori_IMG)
elif i > 98 and i <= 998:
ori_IMG = 'images/00' + str(i+1) + '.jpg'
imgs.append(ori_IMG)
elif i > 998 :
ori_IMG = 'images/0' + str(i+1) + '.jpg'
imgs.append(ori_IMG)
else:
ori_IMG = 'images/0000' + str(i+1) + '.jpg'
imgs.append(ori_IMG)
for i in imgs:
# print i
img = mpimg.imread(i)
img = transform.resize(img, numpy.array([64,64]))
img = rgb2gray(img)
result = hog(img)
result = (result.ravel()).tolist()
results.append(result)
f = open('lables.txt')
data = f.readlines()
f.close
for i in range(len(data)):
data[i] = int(data[i][10])
data = data[0:2000]
clf = svm.LinearSVC()
clf.fit(results,data)
filename = "savedLinearSVM2000.pkl"
with open(filename,'w') as f:
pname = pickle.dump(clf,f)
print "DONE!"
__author__ = 'Wong Sylvia'
from myHog import hog
import cPickle as pickle
from sklearn import svm
import matplotlib.image as mpimg
from skimage.color import rgb2gray
from skimage import transform
import numpy
filename = "savedlinearSVM100.pkl"
with open(filename, 'r') as f:
clf = pickle.load(f)
test1 = mpimg.imread("test1.jpg")
test1 = transform.resize(test1, numpy.array([64, 64]))
test1 = rgb2gray(test1)
test1 = hog(test1).ravel()
test0 = mpimg.imread("test0.jpg")
test0 = transform.resize(test0, numpy.array([64, 64]))
test0 = rgb2gray(test0)
test0 = hog(test0).ravel()
print "Test Picture : Human"
print clf.predict(test1)
print "Test Picture : Non-Human"
print clf.predict(test0)
