【tensorflow】MTCNN网络基本函数P_Net（net_factory）

def P_Net(inputs,label=None,bbox_target=None,landmark_target=None,training=True):
    #define common param
    #关于slim.arg_scope的用法大家可以自行查找，网上有很多关于这个的介绍
    with slim.arg_scope([slim.conv2d],
                        activation_fn=prelu,
                        weights_initializer=slim.xavier_initializer(),
                        biases_initializer=tf.zeros_initializer(),
                        weights_regularizer=slim.l2_regularizer(0.0005), 
                        padding='valid'):
        print(inputs.get_shape())
        #输出input的shape


        net = slim.conv2d(inputs, 10, 3, stride=1,scope='conv1')
        #卷积操作，卷积核的个数是10，卷积核的形式是[3,3]，步长为1，其余的参数和上面的slim.arg_scope一样
        _activation_summary(net)
        #数据的记录（具体的解释看本博客的下一个函数解释）
        print(net.get_shape())
        #输出net的shape
        net = slim.max_pool2d(net, kernel_size=[2,2], stride=2, scope='pool1', padding='SAME')
        #池化
        _activation_summary(net)
        #同上
        print(net.get_shape())
        #同上
        net = slim.conv2d(net,num_outputs=16,kernel_size=[3,3],stride=1,scope='conv2')
        #卷积
        _activation_summary(net)
        #同上
        print(net.get_shape())
        #同上
        net = slim.conv2d(net,num_outputs=32,kernel_size=[3,3],stride=1,scope='conv3')
        #卷积
        _activation_summary(net)
        #同上
        print(net.get_shape())
        #同上
        #batch*H*W*2
        conv4_1 = slim.conv2d(net,num_outputs=2,kernel_size=[1,1],stride=1,scope='conv4_1',activation_fn=tf.nn.softmax)
        #卷积；这里用了softmax函数，是个二分类的问题
        _activation_summary(conv4_1)
        #同上
        #conv4_1 = slim.conv2d(net,num_outputs=1,kernel_size=[1,1],stride=1,scope='conv4_1',activation_fn=tf.nn.sigmoid)
        
        print (conv4_1.get_shape())
        #输出的shape是[batch,1,1,2]
        
        #batch*H*W*4，（bbox的输入shape）
        bbox_pred = slim.conv2d(net,num_outputs=4,kernel_size=[1,1],stride=1,scope='conv4_2',activation_fn=None)
        _activation_summary(bbox_pred)
        print (bbox_pred.get_shape())
        
        #batch*H*W*10（landmark的输入shape）
        landmark_pred = slim.conv2d(net,num_outputs=10,kernel_size=[1,1],stride=1,scope='conv4_3',activation_fn=None)
        _activation_summary(landmark_pred)
        print (landmark_pred.get_shape())


        # add projectors for visualization





        #cls_prob_original = conv4_1 
        #bbox_pred_original = bbox_pred
        #training=True
        if training:
            #batch*2
            # calculate classification loss
            cls_prob = tf.squeeze(conv4_1,[1,2],name='cls_prob')
            #tf.squeeze()删除conv4_1中所指定位置大小是1的维度,shape是[batch,2]
            cls_loss = cls_ohem(cls_prob,label)
            #获得人脸分类训练的loss值
            
            #batch
            # cal bounding box error, squared sum error
            #获得人脸框训练的loss值
            bbox_pred = tf.squeeze(bbox_pred,[1,2],name='bbox_pred')
            bbox_loss = bbox_ohem(bbox_pred,bbox_target,label)
            #batch*10
            #获得关键点训练的loss值
            landmark_pred = tf.squeeze(landmark_pred,[1,2],name="landmark_pred")
            landmark_loss = landmark_ohem(landmark_pred,landmark_target,label)
			#获得正则化损失和分类训练的准确率
            accuracy = cal_accuracy(cls_prob,label)
            L2_loss = tf.add_n(slim.losses.get_regularization_losses())
            return cls_loss,bbox_loss,landmark_loss,L2_loss,accuracy
        #test
        else:
            #when test,batch_size = 1
            cls_pro_test = tf.squeeze(conv4_1, axis=0)
            bbox_pred_test = tf.squeeze(bbox_pred,axis=0)
            landmark_pred_test = tf.squeeze(landmark_pred,axis=0)
            return cls_pro_test,bbox_pred_test,landmark_pred_test