PaddlePaddle动态图实现VGG（眼底筛查为例）

本案例参考课程：百度架构师手把手教深度学习的内容。 主要目的为练习vgg动态图的PaddlePaddle实现。

本案例已经在AISTUDIO共享，链接为：

https://aistudio.baidu.com/aistudio/projectdetail/244766

数据集iChallenge-PM：

数据集图片 iChallenge-PM中既有病理性近视患者的眼底图片，也有非病理性近视患者的图片，命名规则如下：

病理性近视（PM）：文件名以P开头

非病理性近视（non-PM）：

高度近似（high myopia）：文件名以H开头

正常眼睛（normal）：文件名以N开头

我们将病理性患者的图片作为正样本，标签为1； 非病理性患者的图片作为负样本，标签为0。从数据集中选取两张图片，通过LeNet提取特征，构建分类器，对正负样本进行分类，并将图片显示出来。

算法：

VGG VGG是当前最流行的CNN模型之一，2014年由Simonyan和Zisserman提出，其命名来源于论文作者所在的实验室Visual Geometry Group。AlexNet模型通过构造多层网络，取得了较好的效果，但是并没有给出深度神经网络设计的方向。VGG通过使用一系列大小为3x3的小尺寸卷积核和pooling层构造深度卷积神经网络，并取得了较好的效果。VGG模型因为结构简单、应用性极强而广受研究者欢迎，尤其是它的网络结构设计方法，为构建深度神经网络提供了方向。

图3 是VGG-16的网络结构示意图，有13层卷积和3层全连接层。VGG网络的设计严格使用3×33\times 33×3的卷积层和池化层来提取特征，并在网络的最后面使用三层全连接层，将最后一层全连接层的输出作为分类的预测。 在VGG中每层卷积将使用ReLU作为激活函数，在全连接层之后添加dropout来抑制过拟合。使用小的卷积核能够有效地减少参数的个数，使得训练和测试变得更加有效。比如使用两层3×33\times 33×3卷积层，可以得到感受野为5的特征图，而比使用5×55 \times 55×5的卷积层需要更少的参数。由于卷积核比较小，可以堆叠更多的卷积层，加深网络的深度，这对于图像分类任务来说是有利的。VGG模型的成功证明了增加网络的深度，可以更好的学习图像中的特征模式。

关键代码：

import os
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
from PIL import Image

DATADIR = '/home/aistudio/work/palm/PALM-Training400/PALM-Training400'
# 文件名以N开头的是正常眼底图片，以P开头的是病变眼底图片
file1 = 'N0012.jpg'
file2 = 'P0095.jpg'

# 读取图片
img1 = Image.open(os.path.join(DATADIR, file1))
img1 = np.array(img1)
img2 = Image.open(os.path.join(DATADIR, file2))
img2 = np.array(img2)

# 画出读取的图片
plt.figure(figsize=(16, 8))
f = plt.subplot(121)
f.set_title('Normal', fontsize=20)
plt.imshow(img1)
f = plt.subplot(122)
f.set_title('PM', fontsize=20)
plt.imshow(img2)
plt.show()

In[3]

# 查看图片形状
img1.shape, img2.shape

((2056, 2124, 3), (2056, 2124, 3))

In[5]

#定义数据读取器
import cv2
import random
import numpy as np

# 对读入的图像数据进行预处理
def transform_img(img):
    # 将图片尺寸缩放道 224x224
    img = cv2.resize(img, (224, 224))
    # 读入的图像数据格式是[H, W, C]
    # 使用转置操作将其变成[C, H, W]
    img = np.transpose(img, (2,0,1))
    img = img.astype('float32')
    # 将数据范围调整到[-1.0, 1.0]之间
    img = img / 255.
    img = img * 2.0 - 1.0
    return img

# 定义训练集数据读取器
def data_loader(datadir, batch_size=10, mode = 'train'):
    # 将datadir目录下的文件列出来，每条文件都要读入
    filenames = os.listdir(datadir)
    def reader():
        if mode == 'train':
            # 训练时随机打乱数据顺序
            random.shuffle(filenames)
        batch_imgs = []
        batch_labels = []
        for name in filenames:
            filepath = os.path.join(datadir, name)
            img = cv2.imread(filepath)
            img = transform_img(img)
            if name[0] == 'H' or name[0] == 'N':
                # H开头的文件名表示高度近似，N开头的文件名表示正常视力
                # 高度近视和正常视力的样本，都不是病理性的，属于负样本，标签为0
                label = 0
            elif name[0] == 'P':
                # P开头的是病理性近视，属于正样本，标签为1
                label = 1
            else:
                raise('Not excepted file name')
            # 每读取一个样本的数据，就将其放入数据列表中
            batch_imgs.append(img)
            batch_labels.append(label)
            if len(batch_imgs) == batch_size:
                # 当数据列表的长度等于batch_size的时候，
                # 把这些数据当作一个mini-batch，并作为数据生成器的一个输出
                imgs_array = np.array(batch_imgs).astype('float32')
                labels_array = np.array(batch_labels).astype('float32').reshape(-1, 1)
                yield imgs_array, labels_array
                batch_imgs = []
                batch_labels = []

        if len(batch_imgs) > 0:
            # 剩余样本数目不足一个batch_size的数据，一起打包成一个mini-batch
            imgs_array = np.array(batch_imgs).astype('float32')
            labels_array = np.array(batch_labels).astype('float32').reshape(-1, 1)
            yield imgs_array, labels_array

    return reader

# 查看数据形状
DATADIR = '/home/aistudio/work/palm/PALM-Training400/PALM-Training400'
train_loader = data_loader(DATADIR, 
                           batch_size=10, mode='train')
data_reader = train_loader()
data = next(data_reader)
data[0].shape, data[1].shape

((10, 3, 224, 224), (10, 1))

In[6]

!pip install xlrd 

import pandas as pd
df=pd.read_excel('/home/aistudio/work/palm/PALM-Validation-GT/PM_Label_and_Fovea_Location.xlsx')
df.to_csv('/home/aistudio/work/palm/PALM-Validation-GT/labels.csv',index=False)

 

#训练和评估代码
import os
import random
import paddle
import paddle.fluid as fluid
import numpy as np

DATADIR = '/home/aistudio/work/palm/PALM-Training400/PALM-Training400'
DATADIR2 = '/home/aistudio/work/palm/PALM-Validation400'
CSVFILE = '/home/aistudio/work/palm/PALM-Validation-GT/labels.csv'

# 定义训练过程
def train(model):
    with fluid.dygraph.guard():
        print('start training ... ')
        model.train()
        epoch_num = 5
        # 定义优化器
        opt = fluid.optimizer.Momentum(learning_rate=0.001, momentum=0.9)
        # 定义数据读取器，训练数据读取器和验证数据读取器
        train_loader = data_loader(DATADIR, batch_size=10, mode='train')
        valid_loader = valid_data_loader(DATADIR2, CSVFILE)
        for epoch in range(epoch_num):
            for batch_id, data in enumerate(train_loader()):
                x_data, y_data = data
                img = fluid.dygraph.to_variable(x_data)
                label = fluid.dygraph.to_variable(y_data)
                # 运行模型前向计算，得到预测值
                logits = model(img)
                # 进行loss计算
                loss = fluid.layers.sigmoid_cross_entropy_with_logits(logits, label)
                avg_loss = fluid.layers.mean(loss)

                if batch_id % 10 == 0:
                    print("epoch: {}, batch_id: {}, loss is: {}".format(epoch, batch_id, avg_loss.numpy()))
                # 反向传播，更新权重，清除梯度
                avg_loss.backward()
                opt.minimize(avg_loss)
                model.clear_gradients()

            model.eval()
            accuracies = []
            losses = []
            for batch_id, data in enumerate(valid_loader()):
                x_data, y_data = data
                img = fluid.dygraph.to_variable(x_data)
                label = fluid.dygraph.to_variable(y_data)
                # 运行模型前向计算，得到预测值
                logits = model(img)
                # 二分类，sigmoid计算后的结果以0.5为阈值分两个类别
                # 计算sigmoid后的预测概率，进行loss计算
                pred = fluid.layers.sigmoid(logits)
                loss = fluid.layers.sigmoid_cross_entropy_with_logits(logits, label)
                # 计算预测概率小于0.5的类别
                pred2 = pred * (-1.0) + 1.0
                # 得到两个类别的预测概率，并沿第一个维度级联
                pred = fluid.layers.concat([pred2, pred], axis=1)
                acc = fluid.layers.accuracy(pred, fluid.layers.cast(label, dtype='int64'))
                accuracies.append(acc.numpy())
                losses.append(loss.numpy())
            print("[validation] accuracy/loss: {}/{}".format(np.mean(accuracies), np.mean(losses)))
            model.train()

        # save params of model
        fluid.save_dygraph(model.state_dict(), 'mnist')
        # save optimizer state
        fluid.save_dygraph(opt.state_dict(), 'mnist')


# 定义评估过程
def evaluation(model, params_file_path):
    with fluid.dygraph.guard():
        print('start evaluation .......')
        #加载模型参数
        model_state_dict, _ = fluid.load_dygraph(params_file_path)
        model.load_dict(model_state_dict)

        model.eval()
        eval_loader = load_data('eval')

        acc_set = []
        avg_loss_set = []
        for batch_id, data in enumerate(eval_loader()):
            x_data, y_data = data
            img = fluid.dygraph.to_variable(x_data)
            label = fluid.dygraph.to_variable(y_data)
            # 计算预测和精度
            prediction, acc = model(img, label)
            # 计算损失函数值
            loss = fluid.layers.cross_entropy(input=prediction, label=label)
            avg_loss = fluid.layers.mean(loss)
            acc_set.append(float(acc.numpy()))
            avg_loss_set.append(float(avg_loss.numpy()))
        # 求平均精度
        acc_val_mean = np.array(acc_set).mean()
        avg_loss_val_mean = np.array(avg_loss_set).mean()

        print('loss={}, acc={}'.format(avg_loss_val_mean, acc_val_mean))

In[8]

# -*- coding:utf-8 -*-

# VGG模型代码
import numpy as np
import paddle
import paddle.fluid as fluid
from paddle.fluid.layer_helper import LayerHelper
from paddle.fluid.dygraph.nn import Conv2D, Pool2D, BatchNorm, FC
from paddle.fluid.dygraph.base import to_variable

# 定义vgg块，包含多层卷积和1层2x2的最大池化层
class vgg_block(fluid.dygraph.Layer):
    def __init__(self, name_scope, num_convs, num_channels):
        """
        num_convs, 卷积层的数目
        num_channels, 卷积层的输出通道数，在同一个Incepition块内，卷积层输出通道数是一样的
        """
        super(vgg_block, self).__init__(name_scope)
        self.conv_list = []
        for i in range(num_convs):
            conv_layer = self.add_sublayer('conv_' + str(i), Conv2D(self.full_name(), 
                                        num_filters=num_channels, filter_size=3, padding=1, act='relu'))
            self.conv_list.append(conv_layer)
        self.pool = Pool2D(self.full_name(), pool_stride=2, pool_size = 2, pool_type='max')
    def forward(self, x):
        for item in self.conv_list:
            x = item(x)
        return self.pool(x)

class VGG(fluid.dygraph.Layer):
    def __init__(self, name_scope, conv_arch=((2, 64), 
                                (2, 128), (3, 256), (3, 512), (3, 512))):
        super(VGG, self).__init__(name_scope)
        self.vgg_blocks=[]
        iter_id = 0
        # 添加vgg_block
        # 这里一共5个vgg_block，每个block里面的卷积层数目和输出通道数由conv_arch指定
        for (num_convs, num_channels) in conv_arch:
            block = self.add_sublayer('block_' + str(iter_id), 
                    vgg_block(self.full_name(), num_convs, num_channels))
            self.vgg_blocks.append(block)
            iter_id += 1
        self.fc1 = FC(self.full_name(),
                      size=4096,
                      act='relu')
        self.drop1_ratio = 0.5
        self.fc2= FC(self.full_name(),
                      size=4096,
                      act='relu')
        self.drop2_ratio = 0.5
        self.fc3 = FC(self.full_name(),
                      size=1,
                      )
    def forward(self, x):
        for item in self.vgg_blocks:
            x = item(x)
        x = fluid.layers.dropout(self.fc1(x), self.drop1_ratio)
        x = fluid.layers.dropout(self.fc2(x), self.drop2_ratio)
        x = self.fc3(x)
        return x

 

with fluid.dygraph.guard():
    model = VGG("VGG")

train(model)

start training ... 
epoch: 0, batch_id: 0, loss is: [0.7242754]
epoch: 0, batch_id: 10, loss is: [0.6634571]
epoch: 0, batch_id: 20, loss is: [0.7898234]
epoch: 0, batch_id: 30, loss is: [0.60537547]
[validation] accuracy/loss: 0.9424999952316284/0.35623037815093994
epoch: 1, batch_id: 0, loss is: [0.31599292]
epoch: 1, batch_id: 10, loss is: [0.1198744]
epoch: 1, batch_id: 20, loss is: [0.46862125]
epoch: 1, batch_id: 30, loss is: [0.2300901]
[validation] accuracy/loss: 0.92249995470047/0.2342415601015091
epoch: 2, batch_id: 0, loss is: [0.22039299]
epoch: 2, batch_id: 10, loss is: [0.65977865]
epoch: 2, batch_id: 20, loss is: [0.37409317]
epoch: 2, batch_id: 30, loss is: [0.1841044]
[validation] accuracy/loss: 0.9325000643730164/0.22097690403461456
epoch: 3, batch_id: 0, loss is: [0.4992897]
epoch: 3, batch_id: 10, loss is: [0.31177607]
epoch: 3, batch_id: 20, loss is: [0.1721839]
epoch: 3, batch_id: 30, loss is: [0.38319916]
[validation] accuracy/loss: 0.9199999570846558/0.20679759979248047
epoch: 4, batch_id: 0, loss is: [0.20610766]
epoch: 4, batch_id: 10, loss is: [0.06688808]
epoch: 4, batch_id: 20, loss is: [0.3352648]
epoch: 4, batch_id: 30, loss is: [0.28062168]
[validation] accuracy/loss: 0.9149999618530273/0.21788272261619568

 

with fluid.dygraph.guard():
    model = VGG("VGG")

train(model)