def post(self):
image_stream = self.request.files['img'][0]
image_64_encode = base64.encodestring(image_stream['body'])
image_64_decode = base64.decodestring(image_64_encode)
print("-------Detection Started-----")
detected_imgs = prep.init(image_64_decode)
print("-------Detection Complete-----")
print("No. of Detected Images ",len(detected_imgs))
print("-------Recognition Started-----")
model = 'model/lightened_cnn/lightened_cnn'
para = cmp.loadModel(model, 0)
for index in range(len(detected_imgs)):
inputPath = '/Users/admin/Work/Face_Recognition/out'
print("Recognition started for Face ",index+1 )
cmp.compare_two_face(para, inputPath,detected_imgs[index])
print("-------Recognition Complete-----")
def process1(image_urls,d,folder):
df = pd.DataFrame(dtype=np.int8)
hashtable = {}
st = time.time()
for i in range(len(image_urls)):
print("**************" + image_urls[i] + "*************************")
img = urllib.request.urlopen(image_urls[i]).read()
image_id = image_urls[i].split('/')[-1]
image_64_encode = base64.encodestring(img)
image_64_decode = base64.decodestring(image_64_encode)
# image_64_decode = cv2.resize(image_64_decode, (image_64_decode.shape[0] / 4, image_64_decode.shape[1] / 4))
dst = time.time()
detected_imgs = prep.init(image_64_decode)
det = time.time()
print("Time Taken In Detection on No. of faces ", (det - dst), len(detected_imgs))
for index in range(len(detected_imgs)):
print("For Face************", index)
rst = time.time()
detected_face_id = image_id + '_' + str(index)
noOfFacesInImage = len(detected_imgs)
response = cmp.compare_two_face(para, detected_imgs[index], hashtable, detected_face_id, image_id,
noOfFacesInImage)
ret = time.time()
# print("is_already_exist",response)
print("Time Taken In Recognition ", (ret - rst))
if (response['is_already_exist'] == True):
buildMatrix(response['classId'], image_id, noOfFacesInImage, df)
elif (response['is_already_exist'] == False):
buildHashTable(detected_face_id, detected_imgs[index], hashtable)
buildMatrix(detected_face_id, image_id, noOfFacesInImage, df)
# print(hashtable)
df.fillna(0, inplace=True)
print("**********dataFrame************")
print(df)
#print(image_urls[i].split('/')[-1])
name = os.path.join(d,folder+'_matrix.csv')
print(name)
df.to_csv(name)
dff = get_most_frequent_faces(df)
print("**********Most Frequent faces dataFrame************")
print(dff)
name = os.path.join(d, folder + '_result.csv')
dff.to_csv(name)
pics = get_pics_with_most_faces(dff)
for p in pics[0:3]:
os.rename(os.path.join(d,p), os.path.join(d,'top_'+p))
name = os.path.join(d, folder + '_result.csv')
#pics.to_csv(name)
# analyse(df)
print("**********Final Output************")
print(pics)
et = time.time()
logger.info("Time Taken In Detection And Recognition in folder {} {}".format(folder,((et - st) / 60)))
def get(self):
from recognition import compare as cmp
model='model/lightened_cnn/lightened_cnn'
para = cmp.loadModel(model,0)
print ("similar",cmp.compare_two_face(para,"/Users/admin/Work/Face_Recognition/output","151497675637119.jpg","151377336968480.jpg"))
#cmp.compare_two_face(para, "/Users/admin/Work/facerecognition/out", "151426961291577.jpg")
print("detect start")
self.write("recognize")
def detectImageAndAssignClasses(obj, step, callback=None):
image_64_decode = decodeImage(obj)
print(
"***************-------Detection Started FROM CAMERA-----*******************",
step)
detected_imgs = prep.init(image_64_decode)
model = 'model/lightened_cnn/lightened_cnn'
para = cmp.loadModel(model, 0)
for index in range(len(detected_imgs)):
response = cmp.compare_two_face(
obj, para, os.environ.get("DETECTED_FACES_STORE_PATH"),
detected_imgs[index], step, image_64_decode,
createRandomImageId("main_img"))
print(
"***************-------Detection And Recognition Complete FROM CAMERA-----*******************",
step)
callback(response)
def process():
dir = "/Users/admin/Work/jaws-recognition-service/main"
items = os.listdir(dir)
print(items)
i = 0
final = []
resp = []
elements = []
image_urls = []
# Append empty lists in first two indexes.
elements.append([])
for folder in items:
res = {}
# images = os.listdir(os.path.join(dir, folder))
url = 'http://localhost:8888/main/' + folder
image_urls.append(url)
# for img in images:
# url = 'http://localhost:8888/main/' + img
# image_urls.append(url)
# image_urls = ['http://localhost:8888/main/539266179/chetan_ashish.jpg','http://localhost:8888/main/539266232/faraz.jpg','http://localhost:8888/main/539266232/faraj_ashish_rahul.jpg']
print(image_urls)
df = pd.DataFrame(dtype=np.int8)
hashtable = {}
st = time.time()
for i in range(len(image_urls)):
img = urllib.request.urlopen(image_urls[i]).read()
image_id = image_urls[i].split('/')[-1]
image_64_encode = base64.encodestring(img)
image_64_decode = base64.decodestring(image_64_encode)
dst = time.time()
detected_imgs = prep.init(image_64_decode)
det =time.time()
print("Time Taken In Detection on No. of faces ", (det - dst),len(detected_imgs))
for index in range(len(detected_imgs)):
print("For Face************",index)
rst = time.time()
detected_face_id =image_id + '_' + str(index)
noOfFacesInImage = len(detected_imgs)
response = cmp.compare_two_face(para,detected_imgs[index],hashtable,detected_face_id,image_id,noOfFacesInImage)
ret = time.time()
#print("is_already_exist",response)
print("Time Taken In Recognition ", (ret - rst))
if(response['is_already_exist'] == True):
buildMatrix(response['classId'], image_id, noOfFacesInImage,df)
elif(response['is_already_exist'] == False):
buildHashTable(detected_face_id,detected_imgs[index],hashtable)
buildMatrix(detected_face_id,image_id,noOfFacesInImage,df)
#print(hashtable)
df.fillna(0, inplace=True)
print("**********dataFrame************")
print(df)
dff = get_most_frequent_faces(df)
print("**********Most Frequent faces dataFrame************")
print(dff)
pics = get_pics_with_most_faces(dff)
#analyse(df)
print("**********Final Output************")
print(pics)
et = time.time()
print("Final Time Taken In Detection And Recognition ", (et - st))