MobileNet-V2

MobileNet-V2

背景介绍

  MobileNet-V2:是MobileNet的升级版本，Google公司于2018年提出，是一个结构简单，参数量较少的轻量级深度学习网络模型。

MobileNet-V2特点

  保留MobileNet的SeparableConv深度可分离卷积结构，并且在深度可分离卷积前通过1x1卷积提升通道数。
  引入了反残差结构，借鉴了ResNet的思想，联系了不同尺度的特征
  引入了线性瓶颈结构，在残差相加前不使用ReLU6激活函数，避免破坏特征

Separable Convolution

  Separable Convolution(深度可分离卷积)：是上面两个卷积合二为一的卷积操作。
  第一步：DepthwiseConv，对每一个通道进行卷积
  第二步：PointwiseConv，对第一步得到的结果进行1x1卷积，实现通道融合
  主要作用是大大降低网络的参数量，并且可以调整为任意合适的通道数。第一步的目的是减少参数量，第二步是调整通道数，因此将两个卷积操作结合，组成深度可分离卷积。

MobileNet-V2图像分析

TensorFlow2.0实现

from functools import reduce
import tensorflow.keras as keras


def compose(*funcs):
    if funcs:
        return reduce(lambda f, g: lambda *a, **kw: g(f(*a, **kw)), funcs)
    else:
        raise ValueError('Composition of empty sequence not supported.')


class Conv_Bn_Relu6(keras.layers.Layer):
    def __init__(self, filters, kernel_size, strides, padding, name):
        super(Conv_Bn_Relu6, self).__init__()
        self._name = name
        self.block = keras.Sequential()
        if name.find('depthwise') == -1:
            self.block.add(keras.layers.Conv2D(filters, kernel_size, strides, padding=padding))
        else:
            self.block.add(keras.layers.DepthwiseConv2D(kernel_size, strides, padding=padding))
        self.block.add(keras.layers.BatchNormalization())
        if name.find('relu') != -1:
            self.block.add(keras.layers.ReLU(6))

    def call(self, inputs, **kwargs):

        return self.block(inputs)


def block(x, filters, t, strides, name):
    shortcut = x

    x = compose(Conv_Bn_Relu6(t * filters, (1, 1), (1, 1), 'same', name='{}_conv_bn_relu6'.format(name)),
                Conv_Bn_Relu6(None, (3, 3), strides, 'same', name='{}_depthwiseconv_bn_relu6'.format(name)),
                Conv_Bn_Relu6(filters, (1, 1), (1, 1), 'same', name='{}_conv_bn'.format(name)))(x)

    if shortcut.shape[-1] == filters and strides == (1, 1):
        x = keras.layers.Add(name='{}_add'.format(name))([x, shortcut])

    return x


def add_block(x, filters, t, strides, n, name):
    x = block(x, filters, t, strides, name='{}_1'.format(name))
    for i in range(n - 1):
        x = block(x, filters, t, (1, 1), name='{}_{}'.format(name, i + 2))

    return x


def mobilenet_v2(input_shape):
    input_tensor = keras.layers.Input(input_shape, name='input')
    x = input_tensor

    x = compose(keras.layers.ZeroPadding2D((1, 1), name='zeropadding'),
                Conv_Bn_Relu6(32, (3, 3), (2, 2), 'valid', name='conv1'))(x)

    x = add_block(x, 16, 1, (1, 1), 1, 'block1')
    x = add_block(x, 24, 6, (2, 2), 2, 'block2')
    x = add_block(x, 32, 6, (2, 2), 3, 'block3')
    x = add_block(x, 64, 6, (2, 2), 4, 'block4')
    x = add_block(x, 96, 6, (1, 1), 3, 'block5')
    x = add_block(x, 160, 6, (2, 2), 3, 'block6')
    x = add_block(x, 320, 6, (1, 1), 1, 'block7')

    x = compose(Conv_Bn_Relu6(1280, (1, 1), (1, 1), 'same', name='conv2'),
                keras.layers.GlobalAveragePooling2D(name='global_averagepool'),
                keras.layers.Dense(1000, activation='softmax', name='dense'))(x)

    model = keras.Model(input_tensor, x, name='MobileNet-V2')

    return model


if __name__ == '__main__':

    model = mobilenet_v2(input_shape=(224, 224, 3))
    model.build(input_shape=(None, 224, 224, 3))
    model.summary()

MobileNet-V2小结

  MobileNet-V2是一种高效的轻量级深度学习网络，MobileNet-V2模型的参数量和MobileNet几乎相同，都是4M的参数量，因为其结合了不同尺度的特征，因此效果优于MobileNet。