類比于貓狗大戰(zhàn)，利用自己搭建的CNN網(wǎng)絡和已經(jīng)搭建好的VGG16實現(xiàn)花朵識別。

1.導入庫

注：導入的庫可能有的用不到，之前打acm時留下的毛病，別管用不用得到，先寫上再說

import tensorflow as tf from tensorflow import keras from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Dense,Conv2D,Flatten,Dropout,MaxPooling2D from tensorflow.keras.preprocessing.image import ImageDataGenerator import os,PIL import numpy as np import matplotlib.pyplot as plt import pathlib from PIL import ImageFile ImageFile.LOAD_TRUNCATED_IMAGES = True

2.數(shù)據(jù)下載

#數(shù)據(jù)下載 dataset_url = "https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz" dataset_dir = tf.keras.utils.get_file(fname = 'flower_photos',origin=dataset_url,untar=True,cache_dir= 'E:/Deep-Learning/flowers') dataset_dir = pathlib.Path(dataset_dir)

下載之后的文件是這個樣子的。

為了方便處理，博主將數(shù)據(jù)集按照8：2的比例手動（就是不會用代碼）劃分成了訓練集和測試集。
處理之后的文件夾如圖所示：

3.計算數(shù)據(jù)總數(shù)

#將train和test下面的數(shù)據(jù)的文件路徑加載到變量中 train_daisy = os.path.join(dataset_dir,"train","daisy") train_dandelion = os.path.join(dataset_dir,"train","dandelion") train_roses = os.path.join(dataset_dir,"train","roses") train_sunflowers = os.path.join(dataset_dir,"train","sunflowers") train_tulips = os.path.join(dataset_dir,"train","tulips")test_daisy = os.path.join(dataset_dir,"test","daisy") test_dandelion = os.path.join(dataset_dir,"test","dandelion") test_roses = os.path.join(dataset_dir,"test","roses") test_sunflowers = os.path.join(dataset_dir,"test","sunflowers") test_tulips = os.path.join(dataset_dir,"test","tulips") #將訓練集和測試集加載到變量中 train_dir = os.path.join(dataset_dir,"train") test_dir = os.path.join(dataset_dir,"test") #統(tǒng)計訓練集和測試集的數(shù)據(jù)數(shù)目 train_daisy_num = len(os.listdir(train_daisy)) train_dandelion_num = len(os.listdir(train_dandelion)) train_roses_num = len(os.listdir(train_roses)) train_sunflowers_num = len(os.listdir(train_sunflowers)) train_tulips_num = len(os.listdir(train_tulips)) train_all = train_tulips_num+train_daisy_num+train_dandelion_num+train_roses_num+train_sunflowers_numtest_daisy_num = len(os.listdir(test_daisy)) test_dandelion_num = len(os.listdir(test_dandelion)) test_roses_num = len(os.listdir(test_roses)) test_sunflowers_num = len(os.listdir(test_sunflowers)) test_tulips_num = len(os.listdir(test_tulips)) test_all = test_tulips_num+test_daisy_num+test_dandelion_num+test_roses_num+test_sunflowers_num

4.超參數(shù)的設置

batch_size = 32 epochs = 10 height = 180 width = 180

5.數(shù)據(jù)預處理

#歸一化處理 train_generator = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1.0/255) test_generator = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1.0/255) #規(guī)定batch_size的大小，文件路徑，打亂圖片順序，規(guī)定圖片的大小 train_data_gen = train_generator.flow_from_directory(batch_size=batch_size,directory=train_dir,shuffle=True,target_size=(height,width),class_mode="categorical") test_data_gen = test_generator.flow_from_directory(batch_size=batch_size,directory=test_dir,shuffle=True,target_size=(height,width),class_mode="categorical")

6.模型搭建&&模型訓練

模型采用的是三層卷積池化層+Dropout+Flatten+兩層全連接層

model = tf.keras.Sequential([tf.keras.layers.Conv2D(16,3,padding="same",activation="relu",input_shape=(height,width,3)),tf.keras.layers.MaxPooling2D(),tf.keras.layers.Conv2D(32,3,padding="same",activation="relu"),tf.keras.layers.MaxPooling2D(),tf.keras.layers.Conv2D(64,3,padding="same",activation="relu"),tf.keras.layers.MaxPooling2D(),tf.keras.layers.Dropout(0.5),tf.keras.layers.Flatten(),tf.keras.layers.Dense(128,activation="relu"),tf.keras.layers.Dense(5,activation='softmax') ]) model.compile(optimizer="adam",loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),metrics=["acc"]) history = model.fit_generator(train_data_gen,steps_per_epoch=train_all//batch_size,epochs=epochs,validation_data=test_data_gen,validation_steps=test_all//batch_size)

實驗效果如下所示：

出現(xiàn)了過擬合的情況，epochs為50的情況下，運行結(jié)果如下所示：

測試集的準確率相比于epochs為10的情況高了10%左右，但是準確率仍然很低，而且過擬合的情況還是很嚴重，運行時間特別長~

7.數(shù)據(jù)增強

train_generator = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1.0/255,rotation_range=45,#傾斜45°width_shift_range=.15,height_shift_range=.15,horizontal_flip=True,#水平翻轉(zhuǎn)zoom_range=0.5)#隨機放大

實驗結(jié)果如下所示：
epochs為20，過擬合情況得到了改善。

loss: 1.2052 - acc: 0.6962 - val_loss: 1.2502 - val_acc: 0.6534 #epochs為20的時候準確率是65，而epochs為50的時候準確率為66，并沒有很大的改善。

8.遷移學習

利用別人訓練好的VGG16網(wǎng)絡對同樣的數(shù)據(jù)進行訓練。

#引用VGG16模型 conv_base = tf.keras.applications.VGG16(weights='imagenet',include_top = False) #設置為不可訓練 conv_base.trainable = False #搭建模型 model = tf.keras.Sequential() model.add(conv_base) model.add(tf.keras.layers.GlobalAveragePooling2D()) model.add(tf.keras.layers.Dense(128,activation='relu')) model.add(tf.keras.layers.Dense(5,activation='sigmoid'))

訓練結(jié)果如下圖所示：

在epochs只有10的情況下，準確率就達到了80，有了明顯的提升。
努力加油a啊

總結(jié)

以上是生活随笔為你收集整理的深度学习之基于卷积神经网络实现花朵识别的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。