def gen_train_sample(im):
# train.classifierclassifier.load('svm_class.xml')
img = pp.preprocess(im.copy())
# img,rot = pp.skew_correction(img)
hight,width=im.shape
# M = cv2.getRotationMatrix2D((hight/2,width/2),rot-90,1)
# im = cv2.warpAffine(im,M,(width,hight))
# cv2.imwrite('skew correct.png',im)
contours2, hierarchy = cv2.findContours(img,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_NONE)
contours = []
for cnt in contours2:
print (cv2.contourArea(cnt))
if(cv2.contourArea(cnt)>20):
contours.append(cnt)
X = [cv2.contourArea(C) for C in contours]
# print len(contours),len(X)
t=[i for i in range (0,len(contours))]
X,t = zip(*sorted(zip(X,t)))
i=0
for j in t:
x,y,w,h=cv2.boundingRect(contours[j])
box = im[y-1:y+h+1,x-1:x+w+1]
char = pp.preprocess(box.copy())
try:
f = train.find_feature(char)
fu= train.np.array(f,train.np.float32)
# print len(fu)
t = train.classifier.predict(fu)
print t
except IndexError:
t = 0
cv2.imwrite('samp/zsamp47_8_'+str(int(t))+'_'+str(i)+'.png',box)
# cv2.imwrite('./samp/'+str(i)+'.png',box)
i+=1
def gen_train_sample(im):
# train.classifierclassifier.load('svm_class.xml')
img = pp.preprocess(im.copy())
# img,rot = pp.skew_correction(img)
hight,width=im.shape
# M = cv2.getRotationMatrix2D((hight/2,width/2),rot-90,1)
# im = cv2.warpAffine(im,M,(width,hight))
# cv2.imwrite('skew correct.png',im)
contours2, hierarchy = cv2.findContours(img,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_NONE)
contours = []
for cnt in contours2:
print (cv2.contourArea(cnt))
if(cv2.contourArea(cnt)>20):
contours.append(cnt)
X = [cv2.contourArea(C) for C in contours]
# print len(contours),len(X)
t=[i for i in range (0,len(contours))]
X,t = zip(*sorted(zip(X,t)))
i=0
for j in t:
x,y,w,h=cv2.boundingRect(contours[j])
box = im[y-1:y+h+1,x-1:x+w+1]
char = pp.preprocess(box.copy())
try:
f = train.find_feature(char)
fu= train.np.array(f,train.np.float32)
# print len(fu)
t = train.classifier.predict(fu)
print t
except IndexError:
t = 0
cv2.imwrite('samp/zsamp47_8_'+str(int(t))+'_'+str(i)+'.png',box)
# cv2.imwrite('./samp/'+str(i)+'.png',box)
i+=1
def __init__(self,char): global cur_char cur_char = self self.hight,t=char.shape self.data=char self.feature=np.array(train.find_feature(self.data.copy()),np.float32) self.label=train.recognize(self.feature)
def train_svm():
# CV2 SVM
svm_params = dict( kernel_type = cv2.SVM_RBF,
svm_type = cv2.SVM_C_SVC,
C=9.34, gamma=15.68 )
svm=cv2.SVM()
label_list=[]
label_list.append('a')
url='train_images/'
train_set = []
s_list=sorted(os.listdir(url))
label = 0
for i in s_list:
s_list=glob.glob(url+i+'/*.png')
# if(len(s_list)>25):
if(len(s_list)>500):
file=open(url+i+'/utf8',"r")
i_uni=file.read()
i_uni=i_uni[:-1]
label_list.append(i_uni)
label+=1
else:
continue
print str(label),i,label_list[label],len(s_list)
int test=10;
for j in s_list:
if(!test-=1)
break;
img=cv2.imread(j,0)
img=pp.preprocess(img)
f =train.find_feature(img.copy())
# print len(f)
s = [label,f]
train_set.append(s)
f=open('label','w')
for l in label_list:
f.write(l+'\n')
f.close()
shuffle(train_set)
f_list = []
label = []
for t in train_set:
label.append(t[0])
f_list.append(t[1])
# np.savetxt('feature.txt',f_list)
# np.savetxt('label.txt',label)
# samples = np.loadtxt('feature.txt',np.float32)
# responses = np.loadtxt('label.txt',np.float32)
# responses = responses.reshape((responses.size,1))
samples = np.array(f_list,np.float32)
responses = np.array(label,np.float32)
print 'auto training initiated'
print 'please wait.....'
svm.train(samples,responses,params=svm_params)
# svm.train_auto(samples,responses,None,None,params=svm_params)
svm.save("svm_class.xml")
def __init__(self, char):
global cur_char
cur_char = self
self.height, t = char.shape
self.data = char
self.feature = np.array(train.find_feature(self.data.copy()),
np.float32)
self.label = train.recognize(self.feature)
def purify_train():
classifier = cv2.SVM()
classifier.load('svm_class.xml')
g=[]
label_list=get_labellist()
# label_list.append('a')
url='train_images/'
v=open('purify.txt','w')
train_set = []
s_list=sorted(os.listdir(url))
label = 0
for i in s_list:
s_list=glob.glob(url+i+'/*.png')
if(len(s_list)>25):
# file=open(url+i+'/utf8',"r")
# i_uni=file.read()
# i_uni=i_uni[:-1]
# label_list.append(i_uni)
label+=1
else:
# for i in s_list1:
continue
print str(label),i,label_list[label],len(s_list)
for j in s_list:
img=cv2.imread(j,0)
# w=find_vlines(img.copy())
img=pp.preprocess(img)
f =train.find_feature(img.copy())
feature = np.array(f,np.float32)
a = classifier.predict(feature)
if a !=label:
q=j.split('/')
print a,label,int(a)
# print label_list[int(a)],i
v.write(q[2]+'\t'+label_list[int(a)]+' '+str(a)+'\t'+str(label)+' '+i+'\n')
# cv2.imwrite('train_im/'+i+'/'+q[2],im)
# os.rename(j,'train_im/'+i+'/'+q[2])
feature=list(feature)
g.append((feature,label))
# print g
with open('data.txt','w') as outfile:
json.dump(g,outfile)
def purify_train():
classifier = cv2.SVM()
classifier.load('svm_class.xml')
g=[]
label_list=get_labellist()
# label_list.append('a')
url='train_images/'
v=open('purify.txt','w')
train_set = []
s_list=sorted(os.listdir(url))
label = 0
for i in s_list:
s_list=glob.glob(url+i+'/*.png')
if(len(s_list)>25):
# file=open(url+i+'/utf8',"r")
# i_uni=file.read()
# i_uni=i_uni[:-1]
# label_list.append(i_uni)
label+=1
else:
# for i in s_list1:
continue
print str(label),i,label_list[label],len(s_list)
for j in s_list:
img=cv2.imread(j,0)
# w=find_vlines(img.copy())
img=pp.preprocess(img)
f =train.find_feature(img.copy())
feature = np.array(f,np.float32)
a = classifier.predict(feature)
if a !=label:
q=j.split('/')
print a,label,int(a)
# print label_list[int(a)],i
v.write(q[2]+'\t'+label_list[int(a)]+' '+str(a)+'\t'+str(label)+' '+i+'\n')
# cv2.imwrite('train_im/'+i+'/'+q[2],im)
# os.rename(j,'train_im/'+i+'/'+q[2])
feature=list(feature)
g.append((feature,label))
# print g
with open('data.txt','w') as outfile:
json.dump(g,outfile)
file=open(url+i+'/utf8',"r")
i_uni=file.read()
i_uni=i_uni[:-1]
label_list.append(i_uni)
label+=1
else:
continue
print str(label),i,label_list[label],len(s_list)
int test=10;
for j in s_list:
if(!test-=1)
break;
img=cv2.imread(j,0)
img=pp.preprocess(img)
f =train.find_feature(img.copy())
# print len(f)
s = [label,f]
train_set.append(s)
f=open('label','w')
for l in label_list:
f.write(l+'\n')
f.close()
shuffle(train_set)
f_list = []
label = []
for t in train_set:
label.append(t[0])
f_list.append(t[1])
samples = np.array(f_list,np.float32)
