VGG识别眼疾iChallenge-PM

内容都是百度AIstudio的内容，我只是在这里做个笔记，不是原创。

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3×3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3×3卷积层，可以得到感受野为5的特征图，而比使用5×5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。

# -*- coding:utf-8 -*-
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 224x224img = 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 - 1.0return 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0label = 0elif name[0] == 'P':# P开头的是病理性近视，属于正样本，标签为1label = 1else: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_arraybatch_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-batchimgs_array = np.array(batch_imgs).astype('float32')labels_array = np.array(batch_labels).astype('float32').reshape(-1, 1)yield imgs_array, labels_arrayreturn reader# 定义验证集数据读取器
def valid_data_loader(datadir, csvfile, batch_size=10, mode='valid'):# 训练集读取时通过文件名来确定样本标签，验证集则通过csvfile来读取每个图片对应的标签# 请查看解压后的验证集标签数据，观察csvfile文件里面所包含的内容# csvfile文件所包含的内容格式如下，每一行代表一个样本，# 其中第一列是图片id，第二列是文件名，第三列是图片标签，# 第四列和第五列是Fovea的坐标，与分类任务无关# ID,imgName,Label,Fovea_X,Fovea_Y# 1,V0001.jpg,0,1157.74,1019.87# 2,V0002.jpg,1,1285.82,1080.47# 打开包含验证集标签的csvfile，并读入其中的内容filelists = open(csvfile).readlines()def reader():batch_imgs = []batch_labels = []for line in filelists[1:]:line = line.strip().split(',')name = line[1]label = int(line[2])# 根据图片文件名加载图片，并对图像数据作预处理filepath = os.path.join(datadir, name)img = cv2.imread(filepath)img = transform_img(img)# 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_arraybatch_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-batchimgs_array = np.array(batch_imgs).astype('float32')labels_array = np.array(batch_labels).astype('float32').reshape(-1, 1)yield imgs_array, labels_arrayreturn reader# LeNet 识别眼疾图片import os
import random
import paddle
import paddle.fluid as fluid
import numpy as npDATADIR = '/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):use_gpu = Trueplace = fluid.CUDAPlace(0) if use_gpu else fluid.CPUPlace()with fluid.dygraph.guard(place):print('start training ... ')model.train()epoch_num = 5# 定义优化器opt = fluid.optimizer.Momentum(learning_rate=0.001, momentum=0.9, parameter_list=model.parameters())# 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 = dataimg = 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 = dataimg = 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 modelfluid.save_dygraph(model.state_dict(), 'mnist')# save optimizer statefluid.save_dygraph(opt.state_dict(), 'mnist')# 定义评估过程
def evaluation(model, params_file_path):use_gpu = Trueplace = fluid.CUDAPlace(0) if use_gpu else fluid.CPUPlace()with fluid.dygraph.guard(place):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 = dataimg = 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))
# 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, Linear
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, num_convs, in_channels, out_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__()self.conv_list = []for i in range(num_convs):conv_layer = self.add_sublayer('conv_' + str(i), Conv2D(num_channels=in_channels, num_filters=out_channels, filter_size=3, padding=1, act='relu'))self.conv_list.append(conv_layer)in_channels = out_channelsself.pool = Pool2D(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, conv_arch=((2, 64), (2, 128), (3, 256), (3, 512), (3, 512))):super(VGG, self).__init__()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in_channels = [3, 64, 128, 256, 512, 512]for (num_convs, num_channels) in conv_arch:block = self.add_sublayer('block_' + str(iter_id), vgg_block(num_convs, in_channels=in_channels[iter_id], out_channels=num_channels))self.vgg_blocks.append(block)iter_id += 1self.fc1 = Linear(input_dim=512*7*7, output_dim=4096,act='relu')self.drop1_ratio = 0.5self.fc2= Linear(input_dim=4096, output_dim=4096,act='relu')self.drop2_ratio = 0.5self.fc3 = Linear(input_dim=4096, output_dim=1)def forward(self, x):for item in self.vgg_blocks:x = item(x)x = fluid.layers.reshape(x, [x.shape[0], -1])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()train(model)