如何解释cnn结果[关闭]

Question

我是深度学习和神经网络的新手，所以现在在完成 youtube 上的教程后，我尝试运行一个活体人脸检测（训练）的代码，代码如下：

from livenessnet import LivenessNet
from sklearn.preprocessing import LabelEncoder
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix
from keras.preprocessing.image import ImageDataGenerator
from keras.applications.mobilenet import preprocess_input
from keras.callbacks import EarlyStopping
from keras.optimizers import Adam
from keras.utils import np_utils
from imutils import paths
from keras import backend as K
import matplotlib.pyplot as plt
import math
import dlib
import numpy as np
import argparse
import cv2
import os

model_path = 'liveness_model.h5'

# initialize the initial learning rate, batch size, and number of
# epochs to train for
INIT_LR = (1e-5)/4
BS = 32
EPOCHS = 20

# construct the training image generator for data augmentation
train_datagen=ImageDataGenerator()

train_generator=train_datagen.flow_from_directory('../db_faces/train',
                                                 target_size=(64,64),
                                                 color_mode='rgb',
                                                 batch_size=BS,
                                                 class_mode='sparse',
                                                 shuffle=True)

validation_generator=train_datagen.flow_from_directory('../db_faces/test',
                                                 target_size=(64,64),
                                                 color_mode='rgb',
                                                 batch_size=BS,
                                                 class_mode='sparse',
                                                 shuffle=False)
 
labels = (train_generator.class_indices)
print(labels)
# initialize the optimizer and model
adam_opt = Adam(lr = INIT_LR, decay = INIT_LR/EPOCHS)
model = LivenessNet.build(width=64, height=64, depth=3,
    classes=len(labels))

model.summary()

print("[INFO] compiling model...")
#configure the learning process
model.compile(loss="sparse_categorical_crossentropy", optimizer= adam_opt,
    metrics=["accuracy"])

early_stopping = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience = 5 )

step_size_train = train_generator.n//train_generator.batch_size
step_size_validation = validation_generator.samples // validation_generator.batch_size

# train the network
print("[INFO] training network for {} epochs...".format(EPOCHS))
H = model.fit(train_generator,
                   steps_per_epoch=step_size_train,
                   validation_data = validation_generator,
                   validation_steps = step_size_validation,
                   epochs=EPOCHS,
                   callbacks = [early_stopping]
                    )

# save the network to disk
print("[INFO] serializing network to '{}'...".format(model_path))
model.save(model_path)

print("[INFO] Class indices")
labels = (train_generator.class_indices)
print(labels)

# summarize history for accuracy
plt.plot(H.history['accuracy'])
plt.plot(H.history['val_accuracy'])
plt.title('model accuracy')
plt.ylabel('accuracy')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()
# summarize history for loss
plt.plot(H.history['loss'])
plt.plot(H.history['val_loss'])
plt.title('model loss')
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()

Y_pred = model.predict(validation_generator, validation_generator.samples // BS + 1)
y_pred = np.argmax(Y_pred, axis=1)
print('Confusion Matrix')
print(confusion_matrix(validation_generator.classes, y_pred))
print('Classification Report')
target_names = ['Fake', 'Real']
print(classification_report(validation_generator.classes, y_pred))

所以在执行后我得到了这 2 个图表和一个带有混淆矩阵的分类报告：

这是矩阵和分类报告：

Confusion Matrix
[[1436  234]
 [  14 1461]]

Classification Report
              precision    recall  f1-score   support
           0       0.99      0.86      0.92      1670
           1       0.86      0.99      0.92      1475
    accuracy                           0.92      3145
   macro avg       0.93      0.93      0.92      3145
weighted avg       0.93      0.92      0.92      3145

所以我必须得到什么作为结论？这个结果好不好？

Answer 1

您的结果很好，因为您看到 f1-score 范围从 0 到 1它没有从 0 到 1 的范围。

精度是真阳性（您的模型预测为真，实际上也是真的）与总阳性的比率，即真阳性 + 假阳性（假阳性 = 您的模型预测为阳性但实际上不是） 简单来说，它基本上告诉我们，如果它预测为 True，它的预测正确的机会是多少。例如，您的模型对 0 类的精度为 0.99，因此如果它预测 0 类，那么预测正确的可能性为 99%。

召回率是真阳性和真阳性 + 假阴性的比率。对于实际上属于 0 类的所有 0 类，召回率告诉我们你的模型有多少被正确识别为属于 0 类。

这有点令人困惑，但阅读一些文章或观看一些 youtube 视频，你会明白的。