AI ninja project [day 12] 圖片分類(2)

2021 iThome 鐵人賽

DAY 12

AI & Data

AI ninja project系列第 12 篇

13th鐵人賽

wilsonsujames

2021-09-12 11:38:12

2383 瀏覽

分享至

這一篇，我想再參考官網的攻略寫一篇，
不過內容多增加了一些程式上的處理，以及過擬合(Overfitting)時的處理。

參考頁面:https://www.tensorflow.org/tutorials/images/classification?hl=zh_tw

首先，引入模組:

import matplotlib.pyplot as plt
import numpy as np
import os
import PIL
import tensorflow as tf

from tensorflow import keras
from tensorflow.keras import layers
from tensorflow.keras.models import Sequential

下載花朵照片資料集，有五類花朵，總共3670張圖片，放置於五個資料夾，
我們可以印出下載路徑，以及查看圖片數目:

import pathlib
dataset_url = "https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz"
data_dir = tf.keras.utils.get_file('flower_photos', origin=dataset_url, untar=True)
data_dir = pathlib.Path(data_dir)
print(data_dir)

image_count = len(list(data_dir.glob('*/*.jpg')))
print(image_count)

切分訓練集以及測試集，validation_split為切分測試集的比例，
而seed為必須給而且需要為一樣的參數(負責洗牌)

batch_size = 32
img_height = 180
img_width = 180


train_ds = tf.keras.preprocessing.image_dataset_from_directory(
  data_dir,
  validation_split=0.2,
  subset="training",
  seed=456,
  image_size=(img_height, img_width),
  batch_size=batch_size)

val_ds = tf.keras.preprocessing.image_dataset_from_directory(
  data_dir,
  validation_split=0.2,
  subset="validation",
  seed=456,
  image_size=(img_height, img_width),
  batch_size=batch_size)

可以查看標籤的內容有那些花:

class_names = train_ds.class_names
print(class_names)

緩存資料(可以給路徑cache("/path/to/file"))，增加訓練速度，
並且將前處理加入pipline:

AUTOTUNE = tf.data.AUTOTUNE

train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)

建立cnn模型，
可以發現第一層，有先進行前處理，將相素數值都除以255，以進行標準化(機器不用算很巨大的數值)

num_classes = 5

model = Sequential([
  layers.experimental.preprocessing.Rescaling(1./255, input_shape=(img_height, img_width, 3)),
  layers.Conv2D(16, 3, padding='same', activation='relu'),
  layers.MaxPooling2D(),
  layers.Conv2D(32, 3, padding='same', activation='relu'),
  layers.MaxPooling2D(),
  layers.Conv2D(64, 3, padding='same', activation='relu'),
  layers.MaxPooling2D(),
  layers.Flatten(),
  layers.Dense(128, activation='relu'),
  layers.Dense(num_classes)
])

model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

epochs設定為10，進行訓練

epochs=10
history = model.fit(
  train_ds,
  validation_data=val_ds,
  epochs=epochs
)

查看訓練過程:

acc = history.history['accuracy']
val_acc = history.history['val_accuracy']

loss = history.history['loss']
val_loss = history.history['val_loss']

epochs_range = range(epochs)

plt.figure(figsize=(8, 8))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, acc, label='Training Accuracy')
plt.plot(epochs_range, val_acc, label='Validation Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')

plt.subplot(1, 2, 2)
plt.plot(epochs_range, loss, label='Training Loss')
plt.plot(epochs_range, val_loss, label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

可以發現訓練集有隨著時間增加準確度，
但測試集的準確度卡在0.65就不在提升了。
測試集的損失函數反而隨著時間增加了，
代表有過擬合(Overfitting)的現象。

處理方式

我們可以使用扭曲、翻轉、歪斜的方式(假設要辨識前方禁止通行交通號誌，就不適合使用這招)來增加訓練集資料。

data_augmentation = keras.Sequential(
  [
    layers.experimental.preprocessing.RandomFlip("horizontal", 
                                                 input_shape=(img_height, 
                                                              img_width,
                                                              3)),
    layers.experimental.preprocessing.RandomRotation(0.1),
    layers.experimental.preprocessing.RandomZoom(0.1),
  ]
)

另一種方法為我們在模型中，加一層Dropout來調節權重:

layers.Dropout(0.2)

重新建立新的模型:

model = Sequential([
  data_augmentation,
  layers.experimental.preprocessing.Rescaling(1./255),
  layers.Conv2D(16, 3, padding='same', activation='relu'),
  layers.MaxPooling2D(),
  layers.Conv2D(32, 3, padding='same', activation='relu'),
  layers.MaxPooling2D(),
  layers.Conv2D(64, 3, padding='same', activation='relu'),
  layers.MaxPooling2D(),
  layers.Dropout(0.2),
  layers.Flatten(),
  layers.Dense(128, activation='relu'),
  layers.Dense(num_classes)
])

model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

進行訓練:

epochs = 15
history = model.fit(
  train_ds,
  validation_data=val_ds,
  epochs=epochs
)

查看訓練結果:

acc = history.history['accuracy']
val_acc = history.history['val_accuracy']

loss = history.history['loss']
val_loss = history.history['val_loss']

epochs_range = range(epochs)

plt.figure(figsize=(8, 8))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, acc, label='Training Accuracy')
plt.plot(epochs_range, val_acc, label='Validation Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')

plt.subplot(1, 2, 2)
plt.plot(epochs_range, loss, label='Training Loss')
plt.plot(epochs_range, val_loss, label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

我們也可以使用官網提供的照片來進行預測，
可以發現由於模型一開始吃資料的時候有多了batch這個張量，所以用tf.expand_dims來增加維度:

sunflower_url = "https://storage.googleapis.com/download.tensorflow.org/example_images/592px-Red_sunflower.jpg"
sunflower_path = tf.keras.utils.get_file('Red_sunflower', origin=sunflower_url)

img = keras.preprocessing.image.load_img(
    sunflower_path, target_size=(img_height, img_width)
)
img_array = keras.preprocessing.image.img_to_array(img)
img_array = tf.expand_dims(img_array, 0) # Create a batch

predictions = model.predict(img_array)
score = tf.nn.softmax(predictions[0])

print(
    "This image most likely belongs to {} with a {:.2f} percent confidence."
    .format(class_names[np.argmax(score)], 100 * np.max(score))
)