2个功能：
1、能够判断摄像头前是否始终同一个人脸
2、将人脸与vgg-face的明星数据库比对，得到你的明星相

# -*- coding: utf-8 -*-
'''set up the vgg-face model using in week4(the same code)'''
from keras.models import Sequential, Model
from keras.layers import Flatten, Dropout, Activation, Permute
from keras.layers import Conv2D, MaxPooling2D
from keras import backend as K
K.set_image_data_format( 'channels_last' )
def convblock(cdim, nb, bits=3):
    L = []
    for k in range(1,bits+1):
        convname = 'conv'+str(nb)+'_'+str(k)
        L.append( Conv2D(cdim,(3,3),padding='same',activation='relu',name=convname) )
    L.append( MaxPooling2D((2,2), strides=(2,2)) )
    return L

def vgg_face_blank():
    withDO = True 
    if True:
        mdl = Sequential()
        mdl.add( Permute((1,2,3), input_shape=(224,224,3)) )
        for l in convblock(64,1, bits=2):
            mdl.add(l)
        for l in convblock(128,2, bits=2):
            mdl.add(l)
        for l in convblock(256,3, bits=3):
            mdl.add(l)   
        for l in convblock(512,4, bits=3):
            mdl.add(l)   
        for l in convblock(512,5, bits=3):
            mdl.add(l)
        mdl.add( Conv2D(4096,kernel_size=(7,7),activation='relu',name='fc6'))
        if withDO:
            mdl.add(Dropout(0.5))
        mdl.add( Conv2D(4096,kernel_size=(1,1),activation='relu',name='fc7'))
        if withDO:
            mdl.add(Dropout(0.5))
        mdl.add(Conv2D(2622,kernel_size=(1,1),activation='relu',name='fc8')) 
        mdl.add(Flatten())
        mdl.add(Activation('softmax'))
        return mdl
    else:
        raise ValueError('not implemented')
facemodel = vgg_face_blank()

from scipy.io import loadmat
data = loadmat('vgg-face.mat', matlab_compatible=False, struct_as_record=False)
l = data['layers']
description = data['meta'][0,0].classes[0,0].description
def copy_mat_to_keras(kmodel):
    kerasnames = [lr.name for lr in kmodel.layers]
    prmt = (0,1,2,3)
    for i in range(l.shape[1]):
        matname = l[0,i][0,0].name[0]
        if matname in kerasnames:
            kindex = kerasnames.index(matname)
            l_weights = l[0,i][0,0].weights[0,0]
            l_bias = l[0,i][0,0].weights[0,1]
            f_l_weights = l_weights.transpose(prmt)
            assert (f_l_weights.shape == kmodel.layers[kindex].get_weights()[0].shape)
            assert (l_bias.shape[1] == 1)
            assert (l_bias[:,0].shape == kmodel.layers[kindex].get_weights()[1].shape)
            assert (len(kmodel.layers[kindex].get_weights()) == 2)
            kmodel.layers[kindex].set_weights([f_l_weights, l_bias[:,0]])
copy_mat_to_keras(facemodel)
def pred(kmodel, crpimg):
    imarr = np.array(crpimg).astype(np.float32)
    imarr = np.expand_dims(imarr, axis=0)
    out = kmodel.predict(imarr)
    best_index = np.argmax(out, axis=1)[0]
    best_name = description[best_index,0]
    return best_name[0]

featuremodel = Model(inputs=facemodel.layers[0].input, outputs=facemodel.layers[-2].output)


#facemodel.summary()
#featuremodel.summary()
'''
set up real time application
function1:detect your face
function2:judge wheater it is still yourself 
function3:who you look like in the dataset of vgg-face(YouTube Faces dataset)
主要是后2个功能，看视频里的人是不是还是一开始的人(用相似度检测，week4的网络(featuremodel)输出特征值后
和第一张人头做相似度计算)，以及看你和哪个明星相似(完整的week4网络facemodel)，
相似性定义我是根据向量的余弦相似度计算的,有点迷，要是低的话，把这个以下的代码部分，再跑一遍好了
程序开始打开摄像头时，等保存了一定数量的图片后直接按ctrl+c停掉，会显示处理速度，大概每秒输出2张
存在新文件夹my_faces里
'''

import cv2
import dlib
import math
import numpy as np
import time
import os
output_dir = './my_faces'
if not os.path.exists(output_dir):
    os.makedirs(output_dir)

detector = dlib.get_frontal_face_detector()
camera = cv2.VideoCapture(0)
index = 1
def features(featmodel, crpimg):   
    imarr = np.array(crpimg).astype(np.float32)
    imarr = np.expand_dims(imarr, axis=0)
    fvec = featmodel.predict(imarr)[0,:]
    normfvec = math.sqrt(fvec.dot(fvec))
    return fvec/normfvec

def cosin_distance(feature1, feature2):#calculate the similarity of two vectors
    dot_product = 0.0
    normA = 0.0
    normB = 0.0
    for a, b in zip(feature1,feature2):
        dot_product += a * b
        normA += a ** 2
        normB += b ** 2
    if normA == 0.0 or normB == 0.0:
        return None
    else:
        return dot_product / ((normA * normB) ** 0.5)

feature_initial=np.zeros(2622,dtype=np.float32)
feature_initial[0]=1.0
start = time.time()
while index>0:
    try:
        print('Being processed picture %s' % index)
        ret, img = camera.read()
        dets = detector(img, 1)
        if (len(dets) > 0):
            for k,d in enumerate(dets):
                cv2.rectangle(img,(d.left(),d.top()),(d.right(),d.bottom()),(255,255,255))
                cropimg=img[d.left():d.right(),d.top():d.bottom()]
                cropimg=cv2.resize(cropimg,(224,224),interpolation=cv2.INTER_CUBIC)
                #predict who you like in famous stars
                similary_face=pred(facemodel, cropimg) 
                cv2.putText(img,'you like '+similary_face, (50, 90), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 0,0), 2)
                feature=features(featuremodel, cropimg)
                if index==1:
                    cv2.putText(img,'photograph is 100% yourself', (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 0,0), 2)
                    feature_initial=feature
                    time.sleep(5)
                else:
                    #predict wheater it is still you
                    user_similarity=cosin_distance(feature, feature_initial)
                    cv2.putText(img,'photograph is '+str(int(user_similarity*100))+'% yourself', (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 0,0), 2)                                    
                cv2.imwrite("./my_faces/"+str(index)+".jpg",img)
                #cv2.imshow('face',cropimg)
        index += 1
        key = cv2.waitKey(2)
    except KeyboardInterrupt:
        end = time.time()
        print('througnput:',(end-start)/index,'frames/per second')
        break
camera.release()
cv2.destroyAllWindows()