ResNext-50模型进行图像识别

本文为为🔗365天深度学习训练营内部文章

原作者：K同学啊

import numpy as np
from keras.preprocessing.image import ImageDataGenerator
from keras.utils import to_categorical
from keras.models import Sequential
from keras.layers import Input,Dense,Dropout,Conv2D,MaxPool2D,Flatten,GlobalAvgPool2D,concatenate \
    ,BatchNormalization,Activation,Add,ZeroPadding2D,Lambda
from keras.layers import ReLU
from keras.optimizers import Adam
import matplotlib.pyplot as plt
from keras.callbacks import LearningRateScheduler
from keras.models import Model

'''
分组卷积模块
'''
# 定义分组卷积
def grouped_convolution_block(init_x,strides,groups,g_channels):
    group_list = []
    # 分组进行卷积
    for c in range(groups):
        # 分组取出数据
        x = Lambda(lambda x:x[:, :, :, c*g_channels:(c+1)*g_channels])(init_x)
        # 分组进行卷积
        x = Conv2D(filters=g_channels,kernel_size=(3,3),strides=strides,padding='same',use_bias=False)(x)
        # 存入list
        group_list.append(x)
    # 合并list中的数据
    group_merge = concatenate(group_list,axis=3)
    x = BatchNormalization(epsilon=1.001e-5)(group_merge)
    x = ReLU()(x)
    return x

'''
定义残差单元
'''
def block(x,filters,strides=1,groups=32,conv_shortcut=True):
    if conv_shortcut:
        shortcut = Conv2D(filters*2,kernel_size=(1,1),strides=strides,padding='same',use_bias=False)(x)
        # epsilon为BN公式中防止分母为0的值
        shortcut = BatchNormalization(epsilon=1.001e-5)(shortcut)
    else:
        shortcut = x

    # 三层卷积层
    x = Conv2D(filters=filters,kernel_size=(1,1),strides=1,padding='same',use_bias=False)(x)
    x = BatchNormalization(epsilon=1.001e-5)(x)
    x = ReLU()(x)
    # 计算每组的通道数
    g_channels = int(filters / groups)
    # 分组进行卷积
    x = grouped_convolution_block(x,strides,groups,g_channels)

    x = Conv2D(filters=filters*2,kernel_size=(1,1),strides=1,padding='same',use_bias=False)(x)
    x = BatchNormalization(epsilon=1.001e-5)(x)
    x = Add()([x,shortcut])
    x = ReLU()(x)
    return x

'''
堆叠残差单元
'''
def stack(x,filters,blocks,strides,groups=32):
    # 每个stack的第一个block的残差连接都需要使用1*1卷积升维
    x = block(x,filters,strides=strides,groups=groups)
    for i in range(blocks):
        x = block(x,filters,groups=groups,conv_shortcut=False)
    return x

'''
搭建ResNext-50网络
'''
def ResNext50(input_shape,num_classes):
    inputs = Input(shape=input_shape)
    # 填充3圈0，[224,224,3]  -> [230,230,3]
    x = ZeroPadding2D((3,3))(inputs)
    x = Conv2D(filters=64,kernel_size=(7,7),strides=2,padding='valid')(x)
    x = BatchNormalization(epsilon=1.001e-5)(x)
    x = ReLU()(x)
    # 填充1圈0
    x = ZeroPadding2D((1,1))(x)
    x = MaxPool2D(pool_size=(3,3),strides=2,padding='valid')(x)
    # 堆叠残差结构
    x = stack(x,filters=128,blocks=2,strides=1)
    x = stack(x,filters=256,blocks=3,strides=2)
    x = stack(x,filters=512,blocks=5,strides=2)
    x = stack(x,filters=1024,blocks=2,strides=2)
    # 根据特征图大小进行全局平均池化
    x = GlobalAvgPool2D()(x)
    x = Dense(num_classes,activation='softmax')(x)
    # 定义模型
    model = Model(inputs=inputs,outputs=x)
    return model

model = ResNext50(input_shape=(224,224,3),num_classes=4)
model.summary()

ResNeXt-50 相比于传统的深层网络（如 ResNet 和 VGG）有明显的优势，特别是在计算效率和模型性能之间找到了较好的平衡。通过引入 卡尔迪纳利性 的概念，ResNeXt-50 能够在网络深度不增加的情况下显著提升模型的能力，同时保持训练的高效性和泛化能力。它适用于各种计算机视觉任务，尤其是在需要高效和准确的图像分类任务中表现出色。