一、項目背景
基于深度學習的面部表情識別
(Facial-expression Recognition)
數(shù)據集cnn_train.csv包含人類面部表情的圖片的label和feature。
在這里，面部表情識別相當于一個分類問題，共有7個類別。
其中l(wèi)abel包括7種類型表情：
一共有28709個label，即包含28709張表情包。
每一行就是一張表情包4848=2304個像素，相當于4848個灰度值(intensity)(0為黑, 255為白)
本項目同時支持GPU與CPU上運行。
二、數(shù)據預處理
1、標簽與特征分離
對原數(shù)據進行處理，分離后分別保存為cnn_label.csv和cnn_data.csv
# cnn_feature_label.py
# ###一、將原始數(shù)據的label和feature(像素)數(shù)據分離
import pandas as pd
# 源數(shù)據路徑
path = '../datasets/originalData/cnn_train.csv'
# 讀取數(shù)據
df = pd.read_csv(path)
# 提取feature(像素)數(shù)據 和 label數(shù)據
df_x = df[['feature']]
df_y = df[['label']]
# 將feature和label數(shù)據分別寫入兩個數(shù)據集
df_x.to_csv('../datasets/cnn_data.csv', index=False, header=False)
df_y.to_csv('../datasets/cnn_label.csv', index=False, header=False)
2、數(shù)據可視化
對特征進一步處理，也就是將每個數(shù)據行的2304個像素值合成每張48*48的表情圖，最后做成24000張表情包。
# face_view.py
# ###二、數(shù)據可視化,將每個數(shù)據行的2304個像素值合成每張48*48的表情圖。
import cv2
import numpy as np
# 放圖片的路徑
path = '../images'
# 讀取像素數(shù)據
data = np.loadtxt('../datasets/cnn_data.csv')
# 按行取數(shù)據并寫圖
for i in range(data.shape[0]):
face_array = data[i, :].reshape((48, 48)) # reshape
cv2.imwrite(path + '//' + '{}.jpg'.format(i), face_array) # 寫圖片
3、分割訓練集和測試集
Step1：劃分一下訓練集和驗證集。一共有28709張圖片，我取前24000張圖片作為訓練集，其他圖片作為驗證集。新建文件夾cnn_train和cnn_val，將0.jpg到23999.jpg放進文件夾cnn_train，將其他圖片放進文件夾cnn_val.
# cnn_picture_label.py
# ###三、表情圖片和類別標注,
# 1.取前24000張圖片作為訓練集放入cnn_train，其他圖片作為驗證集放入cnn_val
# 2.對每張圖片標記屬于哪一個類別，存放在dataset.csv中，分別在剛剛訓練集和測試集執(zhí)行標記任務。
# #因為cpu訓練太慢，我只取前2000張做訓練，400張做測試！！，手動刪除兩個文件夾重dataset.csv的多余行數(shù)據
import os
import pandas as pd
def data_label(path):
# 讀取label文件
df_label = pd.read_csv('../datasets/cnn_label.csv', header=None)
# 查看該文件夾下所有文件
files_dir = os.listdir(path)
# 存放文件名和標簽的列表
path_list = []
label_list = []
# 遍歷所有文件，取出文件名和對應的標簽分別放入path_list和label_list列表
for file_dir in files_dir:
if os.path.splitext(file_dir)[1] == '.jpg':
path_list.append(file_dir)
index = int(os.path.splitext(file_dir)[0])
label_list.append(df_label.iat[index, 0])
# 將兩個列表寫進dataset.csv文件
path_s = pd.Series(path_list)
label_s = pd.Series(label_list)
df = pd.DataFrame()
df['path'] = path_s
df['label'] = label_s
df.to_csv(path + '\\dataset.csv', index=False, header=False)
def main():
# 指定文件夾路徑
train_path = '../datasets/cnn_train'
val_path = '../datasets/cnn_val'
data_label(train_path)
data_label(val_path)
if __name__ == '__main__':
main()
Step2：對每張圖片標記屬于哪一個類別，存放在dataset.csv中，分別在剛剛訓練集和測試集執(zhí)行標記任務。
Step3：重寫Dataset類，它是Pytorch中圖像數(shù)據集加載的一個基類，需要重寫類來實現(xiàn)加載上面的圖像數(shù)據集
# rewrite_dataset.py
# ###四、重寫類來實現(xiàn)加載上面的圖像數(shù)據集。
import bisect
import warnings
import cv2
import numpy as np
import pandas as pd
import torch
import torch.utils.data as data
class FaceDataset(data.Dataset):
# 初始化
def __init__(self, root):
super(FaceDataset, self).__init__()
self.root = root
df_path = pd.read_csv(root + '\\dataset.csv', header=None, usecols=[0])
df_label = pd.read_csv(root + '\\dataset.csv', header=None, usecols=[1])
self.path = np.array(df_path)[:, 0]
self.label = np.array(df_label)[:, 0]
# 讀取某幅圖片，item為索引號
def __getitem__(self, item):
# 圖像數(shù)據用于訓練，需為tensor類型，label用numpy或list均可
face = cv2.imread(self.root + '\\' + self.path[item])
# 讀取單通道灰度圖
face_gray = cv2.cvtColor(face, cv2.COLOR_BGR2GRAY)
# 直方圖均衡化
face_hist = cv2.equalizeHist(face_gray)
"""
像素值標準化
讀出的數(shù)據是48X48的，而后續(xù)卷積神經網絡中nn.Conv2d()
API所接受的數(shù)據格式是(batch_size, channel, width, height)，
本次圖片通道為1，因此我們要將48X48 reshape為1X48X48。
"""
face_normalized = face_hist.reshape(1, 48, 48) / 255.0
face_tensor = torch.from_numpy(face_normalized)
face_tensor = face_tensor.type('torch.FloatTensor')
# face_tensor = face_tensor.type('torch.cuda.FloatTensor')
label = self.label[item]
return face_tensor, label
# 獲取數(shù)據集樣本個數(shù)
def __len__(self):
return self.path.shape[0]
三、搭建模型
卷積神經網絡模型如下：
# CNN_face.py
# 定義一個CNN模型
"""
inputs(48*48*1) ->
conv(24*24*64) -> conv(12*12*128) -> conv(6*6*256) ->
Dropout -> fc(4096) -> Dropout -> fc(1024) ->
outputs(7)
"""
import torch.nn as nn
# 參數(shù)初始化
def gaussian_weights_init(m):
classname = m.__class__.__name__
# 字符串查找find，找不到返回-1，不等-1即字符串中含有該字符
if classname.find('Conv') != -1:
m.weight.data.normal_(0.0, 0.04)
class FaceCNN(nn.Module):
# 初始化網絡結構
def __init__(self):
super(FaceCNN, self).__init__()
# layer1(conv + relu + pool)
# input:(bitch_size, 1, 48, 48), output(bitch_size, 64, 24, 24)
self.conv1 = nn.Sequential(
nn.Conv2d(1, 64, 3, 1, 1),
nn.BatchNorm2d(num_features=64),
nn.RReLU(inplace=True),
nn.MaxPool2d(kernel_size=2, stride=2)
)
# layer2(conv + relu + pool)
# input:(bitch_size, 64, 24, 24), output(bitch_size, 128, 12, 12)
self.conv2 = nn.Sequential(
nn.Conv2d(64, 128, 3, 1, 1),
nn.BatchNorm2d(num_features=128),
nn.RReLU(inplace=True),
nn.MaxPool2d(kernel_size=2, stride=2)
)
# layer3(conv + relu + pool)
# input: (bitch_size, 128, 12, 12), output: (bitch_size, 256, 6, 6)
self.conv3 = nn.Sequential(
nn.Conv2d(128, 256, 3, 1, 1),
nn.BatchNorm2d(num_features=256),
nn.RReLU(inplace=True),
nn.MaxPool2d(kernel_size=2, stride=2)
)
# 參數(shù)初始化
self.conv1.apply(gaussian_weights_init)
self.conv2.apply(gaussian_weights_init)
self.conv3.apply(gaussian_weights_init)
# 全連接層
self.fc = nn.Sequential(
nn.Dropout(p=0.2),
nn.Linear(256*6*6, 4096),
nn.RReLU(inplace=True),
nn.Dropout(p=0.5),
nn.Linear(4096, 1024),
nn.RReLU(inplace=True),
nn.Linear(1024, 256),
nn.RReLU(inplace=True),
nn.Linear(256, 7)
)
# 向前傳播
def forward(self, x):
x = self.conv1(x)
x = self.conv2(x)
x = self.conv3(x)
x = x.view(x.shape[0], -1) # 數(shù)據扁平化
y = self.fc(x)
return y
四、訓練模型
損失函數(shù)使用交叉熵，優(yōu)化器是隨機梯度下降SGD，其中weight_decay為正則項系數(shù)，每輪訓練打印損失值，每10輪訓練打印準確率。
# train.py
# 定義訓練輪
import torch
import torch.utils.data as data
import torch.nn as nn
import numpy as np
from torch import optim
from models import CNN_face
from dataloader import rewrite_dataset
def train(train_dataset, val_dataset, batch_size, epochs, learning_rate, wt_decay, print_cost=True, isPlot=True):
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# 加載數(shù)據集并分割batch
train_loader = data.DataLoader(train_dataset, batch_size)
# 構建模型
model = CNN_face.FaceCNN()
model.to(device)
# 損失函數(shù)和優(yōu)化器
compute_loss = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=learning_rate, weight_decay=wt_decay)
# 學習率衰減
# scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=10, gamma=0.8)
for epoch in range(epochs):
loss = 0
model.train()
for images, labels in train_loader:
optimizer.zero_grad()
outputs = model.forward(images)
loss = compute_loss(outputs, labels)
loss.backward()
optimizer.step()
# 打印損失值
if print_cost:
print('epoch{}: train_loss:'.format(epoch + 1), loss.item())
# 評估模型準確率
if epoch % 10 == 9:
model.eval()
acc_train = validate(model, train_dataset, batch_size)
acc_val = validate(model, val_dataset, batch_size)
print('acc_train: %.1f %%' % (acc_train * 100))
print('acc_val: %.1f %%' % (acc_val * 100))
return model
# 驗證模型在驗證集上的正確率
def validate(model, dataset, batch_size):
val_loader = data.DataLoader(dataset, batch_size)
result, total = 0.0, 0
for images, labels in val_loader:
pred = model.forward(images)
pred_tmp = pred.cuda().data.cpu().numpy()
pred = np.argmax(pred_tmp.data.numpy(), axis=1)
labels = labels.data.numpy()
result += np.sum((pred == labels))
total += len(images)
acc = result / total
return acc
def main():
train_dataset = rewrite_dataset.FaceDataset(root=r'D:\01 Desktop\JUST_YAN\05 DeepLearning\Facial-expression_Reg\datasets\cnn_train')
val_dataset = rewrite_dataset.FaceDataset(root=r'D:\01 Desktop\JUST_YAN\05 DeepLearning\Facial-expression_Reg\datasets\cnn_val')
model = train(train_dataset, val_dataset, batch_size=128, epochs=100, learning_rate=0.01,
wt_decay=0, print_cost=True, isPlot=True)
torch.save(model, 'model_net.pkl') # 保存模型
if __name__ == '__main__':
main()
五、訓練結果
在超參數(shù)為：batch_size=128, epochs=100, learning_rate=0.01, wt_decay=0，的情況下跑得最終結果如下：
文章來源：http://www.zghlxwxcb.cn/news/detail-441954.html
附錄
代碼已托管到GitHub和Gitee：
GitHub：https://github.com/HaoliangZhou/FERNet
Gitee： https://gitee.com/zhou-zhou123c/FERNet
參考資料：
https://blog.csdn.net/Charzous/article/details/107452464/
數(shù)據集
數(shù)據集cnn_train.csv文章來源地址http://www.zghlxwxcb.cn/news/detail-441954.html
到了這里，關于Pytorch實現(xiàn)基于深度學習的面部表情識別（最新，非常詳細）的文章就介紹完了。如果您還想了解更多內容，請在右上角搜索TOY模板網以前的文章或繼續(xù)瀏覽下面的相關文章，希望大家以后多多支持TOY模板網！
![[深度學習實戰(zhàn)]基于PyTorch的深度學習實戰(zhàn)(下)[Mnist手寫數(shù)字圖像識別]](https://imgs.yssmx.com/Uploads/2024/02/603063-1.png)
