paddlepaddle 口罩识别

padddlepaddle口罩识别，查看数据集
导入需要的包

import os
import zipfile
import random
import json
import paddle
import sys
import numpy as np
from PIL import Image
from PIL import ImageEnhance
import paddle.fluid as fluid
from multiprocessing import cpu_count
import matplotlib.pyplot as plt

参数配置

'''
参数配置
'''
train_parameters = {
    "input_size": [3, 224, 224],                              #输入图片的shape
    "class_dim": -1,                                          #分类数
    "src_path":"/home/aistudio/data/data22392/maskDetect.zip",          #原始数据集路径
    "target_path":"/home/aistudio/data/",                     #要解压的路径
    "train_list_path": "/home/aistudio/data/train.txt",       #train.txt路径
    "eval_list_path": "/home/aistudio/data/eval.txt",         #eval.txt路径
    "readme_path": "/home/aistudio/data/readme.json",         #readme.json路径
    "label_dict":{},                                          #标签字典
    "train_batch_size":20,                                    #训练大小
}

数据准备

1.解压数据集

def unzip_data(src_path,target_path):
	'''
	解压原始数据集，将src_path路径下的zip包解压至data目录下
	'''
    if(not os.path.isdir(target_path+"maskDetect")):
        z = zipfile.ZipFile(src_path,'r')
        z.extractall(path=target_path)
        z.close()

def get_data_list(target_path,train_list_path,eval_list_path):

    '''
    生成数据列表
    '''
	#存放所有类别信息
    class_detail = []
	#获取所有类别保存的文件夹名称
    data_list_path = target_path + "maskDetect/"
    class_dirs = os.listdir(data_list_path)
	#总的图像数量
    all_class_images = 0
	#存放类别标签
    class_label = 0
	#存放类别数目
    class_dim = 0
	#存储要写进eval.txt和train.txt中的内容
    trainer_list = []
    eval_list = []
	#读取每个类别，['maskimages','nomaskimages']
    for class_dir in class_dirs:
        if class_dir != ".DS_Store":

            class_dim += 1;
            #每个类别的信息
            class_detail_list = {}
            eval_sum = 0
            trainer_sum = 0
			#统计每个类别有多少张图片
            class_sum = 0
			#获取类别路径
            path = data_list_path + class_dir
			#获取所有图片
            img_paths = os.listdir(path)

            for img_path in img_paths:  #遍历文件夹下每张图片
                name_path = path + '/' + img_path    #每张图片的路径
                if class_sum % 10 == 0:  #每10张图片取一个做验证数据
                    eval_sum += 1
                    eval_list.append(name_path + "\t%d" % class_label + '\n') #验证数据加1
                
                else:
                    trainer_sum += 1
                    trainer_list.append(name_path + "\t%d" % class_label + "\n") #训练数据加1
                
                class_sum += 1 #每类图片的数目
                all_class_images += 1 #所有类图片的数目

 			#说明json文件的class_detail数据           class_detail_list['class_name'] = class_dir #类别名称
            class_detail_list['class_label'] = class_label #类别标签
            class_detail_list['class_eval_images'] = eval_sum  #该类数据的测试集数目
            class_detail_list['class_trainer_images'] = trainer_sum #该类数据的训练集数目
            class_detail.append(class_detail_list)

 #初始化标签列表           train_parameters['label_dict'][str(class_label)] = class_dir
            class_label += 1
	#初始化分类数
    train_parameters['class_dim'] = class_dim

	#打乱数据
    random.shuffle(eval_list)
    with open(eval_list_path,'a') as f:
        for eval_image in eval_list:
            f.write(eval_image)

    random.shuffle(trainer_list)
    with open(train_list_path,'a') as f2:
        for train_image in trainer_list:
            f2.write(train_image)
	#说明json文件信息
    readjson = {}
    readjson['all_class_name'] = data_list_path #文件父目录
    readjson['all_class_images'] = all_class_images
    readjson['class_detail'] = class_detail

    jsons = json.dumps(readjson,sort_keys=True,indent=4,separators=(',', ': '))
    with open(train_parameters['readme_path'],'w') as f:
        f.write(jsons)
    print ('生成数据列表完成！')

'''
参数初始化
'''
src_path=train_parameters['src_path']
target_path=train_parameters['target_path']
train_list_path=train_parameters['train_list_path']
eval_list_path=train_parameters['eval_list_path']
batch_size=train_parameters['train_batch_size']

'''
解压原始数据到指定路径
'''
unzip_data(src_path,target_path)

'''
划分训练集与验证集，乱序，生成数据列表
'''
#每次生成数据列表前，首先清空train.txt和eval.txt
with open(train_list_path, 'w') as f: 
    f.seek(0)
    f.truncate() 
with open(eval_list_path, 'w') as f: 
    f.seek(0)
    f.truncate() 

#生成数据列表   
get_data_list(target_path,train_list_path,eval_list_path)

'''
自定义数据集
'''

import paddle
from paddle.io import Dateset

class MyDataset(paddle.io.Dateset):
	'''
	步骤一：集成paddle.io.Dataset
	'''
	def __init__(self,mode='train'):
		'''
		步骤二：实现构造函数，定义数据集大小
		'''
		super(MyDataset,self).__init__()
		self.data = []
		self.label = []
		if mode == 'train':
			#遍历数据文件
			with open(train_list_path,'r') as f:
			lines = [line.strip() for line in f]
			for line in lines:
				#读入图片文件地址和标签
				img_path, lab = line.strip().split('\t')
				#读入图片
				img = Image.open(img_path)
				#判断图片格式
				if img.mode != "RGB":
					#如果图片格式不是RGB则改为RGB
					img = img.convert('RGB')
				#将图片大小改为(224,224)
				img = img.resize((224,224),Image.BILINEAR)
				#将图片转化为数组
				img  = np.array(img).astype('float32')
				#HWC to CHW
				img = img.transpose((2,0,1))
				#像素值归一化
				img = img/255
				#将数据统一添加到data和label中
				self.data.append(img)
				self.label.append(np.array(lab).astype('int64'))
		else:
			#测试集同上
			with open(eval_list_path,'r') as f:
			lines = [line.strip() for line in f]
			for line in lines:
				img_path,lab = line.strip().split('\t')
				img = Image.open(img_path)
				if img.mode != "RGB":
					img = img.convert('RGB')
				img = img.resize((224,224),Image.BILINEAR)
				img = np.array(img).astypr('float32')
				img = img.transpose((2,0,1))
				img = img / 255
				self.data.append(img)
				self.label.append(bp.array(lab).astype('int64'))
	def __getitem__(self,index):
		 """
	     步骤三：实现__getitem__方法，定义指定index时如何获取数据，并返回单条数据（训练数据，对应的标签）
	     """
	     #返回单一的数据和标签
	     data = self.data[index]
	     label = self.label[index]
	     #注：返回标签数据时必须是int64
	     return data,np.array(label,dtype='int64')
     def __len__(self):
     	return len(self.data) 

#测试定义的数据集
train_dataset = MyDataset(mode='train')

eval_dataset = MyDataset(mode='val')
print('============train_dataset========)

#输出数据集的形状和标签
print(train_dataset.__getitem__(1)[0].shape,train_dataset.__getitem__(1)[1])

#输出数据集长度
print(train_dataset.__len__())

模型配置

VGG的核心是五组卷积操作，每两组之间做Max-Pooling空间降维。同一组内采用多次连续的3x3卷积，卷积核的数目由较浅组的64增多到最深组的512，同一组的卷积核数目是一样的。卷积之后接两层全连接层，之后是分类层。由于每组内卷积层的不同，有11,13,16,19层这几种模型，上图显示一个16层的结构。

class ConvPool(paddle.nn.Layer):
	'''
	卷积+池化
	'''
	def __init__(self,num_channels,
				num_filters,filter_size,pool_size,pool_stride,groups,conv_stride=1,conv_padding=1,act=None,pool_type='max'):
		super(ConvPool,self).__init__()
		self._conv2d_list = []

		for i in range(groups):
			conv2d = self.add_sublayer(#返回一个由所有子层组成的列表。
			'bb_%d' % i,
			paddle.nn.Conv2D(
			in_channels = num_channels, #通道数
			out_channels = num_filters,#卷积核个数
			kernel_size = filter_size,#卷积核大小
			stride = conv_stride,
			padding = conv_padding
			)
		)
			num_channels = num_filters
			self._conv2d_list.append(conv2d)
	
		self._pool2d = paddle.fluid.dygraph.Pool2D(
			pool_size = pool_size, #池化核大小
			pool_type = pool_type,
			pool_stride=pool_stride)

	def forward(self,inputs):
		x = inputs
		for conv in self._conv2d_list:
			x = conv(x)
		x = self._pool2d(x)
		return x

class VGGNet(paddle.nn.Layer):
	'''
	VGG网络
	'''
	def __init__(self):
		super(VGGNet,self).__init__()
		self.convpool01 = ConvPool(
			3,64,3,2,2,2,act='relu')
		self.convpool02 = ConvPool(
			64,128,3,2,2,2,act='relu')
		self.convpool03 = ConvPool(
			128,256,3,2,2,3,act='relu')
		self.convpool04 = ConvPool(
			256,512,3,2,2,3,act = 'relu')
		self.convpool05 = ConvPool(
			512,512,3,2,2,3,act='relu')
		self.pool_5_shape = 512*7*7
		self.fc01 = paddle.nn.Linear(in_features=self.pool_5_shape,out_features=4096)
		self.fc02 = paddle.nn.Linear(
			in_features=4096,out_features=4096)
		self.fc03 = paddle.nn.Linear(in_features=4096,out_features=2)
	def forward(self,inputs):
		#print(intputs.shape) #[8,3,224,224]
		"""前向计算"""
		out = self.convpool01(inputs)
		#print(out.shape) [8,64,112,112]
		out = self.convpool02(out)
		#print(out.shape) [8,128,56,56]
		out = self.convpool03(out)
		#print(out.shape) [8,256,28,28]
		out = self.convpool04(out)
		#print(out.shape) [8,512,14,14]
		out = self.convpool05(out)
		#print(out.shape) [8,512,7,7]
		
		out = paddle.fluid.layers.nn.reshape(out,sshape=[-1,512*7*7])
		out = self.fc01(out)
		out = self.paddle.nn.functional.relu(out)
		out = self.fc02(out)
		out = paddle.nn.functional.relu(out)
		out = self.fc03(out)
		out = paddle.nn.functional.softmax(out)
return out

模型训练与评估

model = paddle.Model(VGGNet())

model.prepare(paddle.optimizer.Adam(parameters=model.parameters()),paddle.nn.CrossEntropyLoss(),paddle.metric.Accuracy(topk=(1,5)))

model.fit(train_dataset,eval_dataset,epochs = 5,batch_size = batch_size,shuffle=True,verbose=1)

模型预测

def load_image(img_path):
    '''
    预测图片预处理
    '''
    #具体过程与自定义数据集差不多，不再赘述
    img = Image.open(img_path) 
    if img.mode != 'RGB': 
        img = img.convert('RGB') 
    img = img.resize((224, 224), Image.BILINEAR)
    img = np.array(img).astype('float32') 
    img = img.transpose((2, 0, 1)) 
    img = img/255               
    return img
label_dic = train_parameters['label_dict']

'''
模型预测
'''
#展示预测图片
infer_path='/home/aistudio/data/data23615/infer_mask01.jpg'
img = Image.open(infer_path)
#绘制图像
plt.imshow(img)         
plt.show()               
#对预测图片进行预处理
infer_path='/home/aistudio/data/data23615/infer_mask01.jpg'
infer_imgs = []
infer_imgs.append(load_image(infer_path))
infer_imgs = np.array(infer_imgs)

for  i in range(len(infer_imgs)):
        data = infer_imgs[i]
        dy_x_data = np.array(data).astype('float32')
        #对读入的图片数据进行扩充维度，使其符合网络输入要求
        dy_x_data = dy_x_data[np.newaxis,np.newaxis,:, : ,:]
        print(dy_x_data.shape)
        #开始预测
        out = model.predict(dy_x_data)
        #输出预测结果
        lab = np.argmax(out[0][0])  
        print("第{}个样本,被预测为：{}".format(i+1,label_dic[str(lab)]))
        
print("结束")