def auto_fix(im, noise_removal=False):
im = auto_crop_hsv(im)
im = auto_crop_hsv(im, crop_white=True)
min_face_size = (int(im.shape[0] * 0.05), int(im.shape[1] * 0.05))
faces = face_detect.detect_faces(im,
min_neighbors=5,
min_size=min_face_size,
max_size=None)
# keep faces above the fold
faces = [face for face in faces if face[1] < im.shape[0] * 0.4]
# find dresses
dresses = [dress_box2(face, im.shape[:2]) for face in faces]
# if len(dresses) > 0:
# print('grabcut!')
# im = grabCut(im, bounding_box(dresses+faces))
if len(faces) > 0:
im = crop_to_human(im, faces, dresses)
# limit max size (after cropping)
im = fit_in(im, 1800, 1200)
if noise_removal:
im = cv2.fastNlMeansDenoisingColored(im)
im = face_detect.skin_detect2(im, marks=True)
# im = simplest_color_balance(im)
# print('retinex starting...')
# im = colorcorrect.algorithm.retinex_with_adjust(im)
# print('retinex complete.')
#face_detect.draw_boxes(im, faces)
#face_detect.draw_boxes(im, dresses, (255, 0, 0))
# face_detect.draw_boxes(im, people, (255, 0, 0))
return im, faces
def auto_fix(im, noise_removal=False):
im = auto_crop_hsv(im)
im = auto_crop_hsv(im, crop_white=True)
min_face_size = (int(im.shape[0]*0.05), int(im.shape[1]*0.05))
faces = face_detect.detect_faces(im, min_neighbors=5, min_size=min_face_size, max_size=None)
# keep faces above the fold
faces = [face for face in faces if face[1] < im.shape[0]*0.4]
# find dresses
dresses = [dress_box2(face, im.shape[:2]) for face in faces]
# if len(dresses) > 0:
# print('grabcut!')
# im = grabCut(im, bounding_box(dresses+faces))
if len(faces) > 0:
im = crop_to_human(im, faces, dresses)
# limit max size (after cropping)
im = fit_in(im, 1800, 1200)
if noise_removal:
im = cv2.fastNlMeansDenoisingColored(im)
im = face_detect.skin_detect2(im, marks=True)
# im = simplest_color_balance(im)
# print('retinex starting...')
# im = colorcorrect.algorithm.retinex_with_adjust(im)
# print('retinex complete.')
#face_detect.draw_boxes(im, faces)
#face_detect.draw_boxes(im, dresses, (255, 0, 0))
# face_detect.draw_boxes(im, people, (255, 0, 0))
return im, faces
def detect_text(image_file):
unlike = ['OF INDIA','INCOME TAX DEPARTMENT','Permanent Account Number','s Name','INDIA','OF INDIA','Birth']
with open(image_file, 'rb') as image:
base64_image = base64.b64encode(image.read()).decode()
url = 'https://vision.googleapis.com/v1/images:annotate?key=AIzaSyAOztXTencncNtoRENa1E3I0jdgTR7IfL0'
header = {'Content-Type': 'application/json'}
body = {
'requests': [{
'image': {
'content': base64_image,
},
'features': [{
'type': 'DOCUMENT_TEXT_DETECTION',
'maxResults': 100,
}]
}]
}
response = requests.post(url, headers=header, json=body).json()
text = response['responses'][0]['textAnnotations'][0]['description'] if len(response['responses'][0]) > 0 else ''
block=str(text).split('\n')
bca=re.findall(r'\s([a-zA-Z]{5}\d{4}[a-zA-Z0-9]{1})',text)
noun = re.compile('([a-zA-Z]+ [a-zA-Z]+ [a-zA-Z]+|[a-zA-Z]+ [a-zA-Z]+|[a-zA-Z]+ [a-zA-Z]+ [a-zA-Z]+ [a-zA-Z]+)')
names = noun.findall(text)
DOB = re.findall(r'\s(\d{2}\/\d{2}\/\d{4})',text)
names = [x for x in names if x not in unlike]
face=detect_faces(image_file,bca[0])
image_string =''
if os.path.isfile(face) is True:
with open(face, 'rb') as image:
image_string = base64.b64encode(image.read()).decode()
os.remove(face)
parsed_birth=str(dateparser.parse(DOB[0],settings={'DATE_ORDER': 'DMY'}).date())
details = {"name":names[0],"father_name":names[1],"pan_no":bca[0],"date_of_birth":parsed_birth,"pan_face":image_string}
return details
def recognize_image(image, model_path, classify=False):
faceRects = detect_faces(image)
faces = crop_rects(image, faceRects)
if classify:
print('Using 1 model to predict')
model = load_model(model_path + 'classification_model.h5')
names = classify_faces(faces, model)
else:
print('Using 7 models to predict')
models = load_models(model_path)
names = predict_faces(faces, models)
window_name = "Recognizing faces"
show_bounding_boxes_and_labels(image, faceRects, names, window_name)
def post(self):
#file = request.form
#base=file['Passport_Image']
image = request.files['Passport_Image']
image_string = base64.b64encode(image.read()).decode()
base = image_string.encode()
details = detect_text(image_string)
print("......................:", details)
strps_base = image_string.strip('')[1]
# print("=============",image_string)
# directory = home +'/'+ date
#print("-------------------------",directory)
#if not os.path.exists(directory):
#print ("path doesn't exist. trying to make")
#os.mkdir(directory)
# name = details['Passport_Document_No'] + '_' +details['FamilyName']
#filename = directory+'/'+name+'.jpeg' # I assume you have a way of picking unique filenames
# t = TemporaryFile(filename)
base = image_string.encode()
print = ("tnkdfkn:", base)
filename = home + '/' + 'document.jpeg' # I assume you have a way of picking unique filenames
with open(filename, 'wb') as f:
f.write(base)
# fullimage_size = ('{:,.0f}'.format(os.path.getsize(filename)/float(1<<10))+" KB")
face = detect_faces(filename)
#os.remove(filename)#print("data return",face)
with open(face, 'rb') as image:
image_string_face = base64.b64encode(image.read()).decode()
#print("image_string:",image_string)
logger.info("Data added successfully to passport")
details['face'] = image_string_face
#logger.info("Data added successfully to passport")
return details
def post(self):
try:
startlog.info("api call hits")
file = request.form
base = file['Passport_Image']
imgdata = base64.b64decode(base)
header = file['scan_type']
details = detect_text(base)
startlog.info(details)
unique_no = str(datetime.now())
if details['type'] == 'PASSPORT':
no = details['data']['Passport_Document_No']
unique_no = no[5:]
elif details['type'] == 'VISA':
no = details['data']['Visa_Number']
unique_no = no[5:]
# elif details['type']=='partial data':
# rand_int=random.randint(0,10000)
# unique_no = rand_int
# I assume you have a way of picking unique filenames
filename = os.path.join(
home, 'xxxx'+str(unique_no)+'document.jpeg')
with open(filename, 'wb') as f:
f.write(imgdata)
if header == 'mobile':
crop = rotate(filename, unique_no)
fullimage_size = ('{:,.0f}'.format(
os.path.getsize(filename)/float(1 << 10))+" KB")
face = detect_faces(crop, unique_no)
os.remove(crop)
os.remove(filename)
elif header == 'web':
fullimage_size = ('{:,.0f}'.format(
os.path.getsize(filename)/float(1 << 10))+" KB")
face = detect_faces(filename, unique_no)
os.remove(filename)
image_string = ''
faceimage_size = ''
if os.path.isfile(face) is True:
with open(face, 'rb') as image:
image_string = base64.b64encode(image.read()).decode()
faceimage_size = ('{:,.0f}'.format(
os.path.getsize(face)/float(1 << 10))+" KB")
os.remove(face)
logger.info("Data added successfully to passport")
details['face'] = image_string
details['fullimage_size'] = fullimage_size
details['faceimage_size'] = faceimage_size
return ({"success": True, "details": details})
except OSError as e:
logger.warning(traceback.format_exc())
return ({"error": str(e), "success": False})
except IndexError as e:
logger.warning(traceback.format_exc())
return ({"message": str(e), "success": False})
except Exception as e:
logger.warning(traceback.format_exc())
return ({"success": False, "message": str(traceback.format_exc())})
import time
import threading
import queue
start = time.time()
que = queue.Queue()
a = queue.Queue()
b = queue.Queue()
c = queue.Queue()
threads_list = list()
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
image = cv2.imread('images/real_1.jpg')
img_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
#Thread 1
t1 = threading.Thread(
target=lambda q, arg1, arg2: q.put(face_detect.detect_faces(arg1, arg2)),
args=(que, face_cascade, img_gray))
t1.start()
t1.join()
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
align1 = que.get()
detections = detector(align1, 1)
faces = dlib.full_object_detections()
for det in detections:
faces.append(predictor(align1, det))
right_eye = [[face.part(i) for i in range(36, 42)] for face in faces]
right_eye = [[(i.x, i.y) for i in eye]
for eye in right_eye] # Convert out of dlib format
right_eye = right_eye[0][0]
right_eye_x = right_eye[0]
import cv2
import numpy as np
import face_detect
face_cascade = cv2.CascadeClassifier("haarcascade_frontalface_default.xml")
img1 = cv2.imread("images/real_3.jpg")
# Converting the image into grayscale
gray=cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
img = face_detect.detect_faces(face_cascade, gray)
cv2.imshow('aligned image',img)
cv2.waitKey(0)
cv2.destroyAllWindows()