LeNET(Mask 데이터 실습)

1-1. 데이터 읽기

• 캐글: https://www.kaggle.com/ashishjangra27/face-mask-12k-images-dataset

# 모듈
import numpy as np
import pandas as pd
import os
import glob
import matplotlib.pyplot as plt
import seaborn as sns
import tensorflow as tf
from tensorflow.keras import Sequential, models
from tensorflow.keras.layers import Flatten, Dense, Conv2D, MaxPool2D
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix

from google.colab import drive
drive.mount('/content/drive')

import zipfile

# 파일 압축 해제
contetn_zip = zipfile.ZipFile('/content/drive/MyDrive/Colab Notebooks/data/archive.zip')
contetn_zip.extractall('./data')
# 객체 종료
contetn_zip.close()

1-2. 파일 정리 후 데이터 프레임

path = './data/Face Mask Dataset/'
dataset = {'image_path': [], 'mask_status':[], 'where':[]}

for where in os.listdir(path):
  for status in os.listdir(path + '/' + where):
    for image in glob.glob(path + where + '/' + status + '/' + '*.png'):
      dataset['image_path'].append(image)
      dataset['mask_status'].append(status)
      dataset['where'].append(where)
      
dataset = pd.DataFrame(dataset)

2. 데이터 나누기

train_df = dataset[dataset['where'] == 'Train']
test_df = dataset[dataset['where'] == 'Test']
valid_df = dataset[dataset['where'] == 'Validation']

# 인덱스 정리
train_df = train_df.reset_index().drop('index', axis=1)

3. 데이터 전처리

data = []
image_size = 150 # 150 x 150

for i in range(len(train_df)):
  # converting the image into grayscale
  img_array = cv2.imread(train_df['image_path'][i], cv2.IMREAD_GRAYSCALE)

  # resizing the array
  new_image_array = cv2.resize(img_array, (image_size, image_size))

  # encoding the image with the label
  # mask o 1 / mask x 0
  if train_df['mask_status'][i] == 'WithMask':
    data.append([new_image_array, 1])
  else:
    data.append([new_image_array, 0])

# 데이터 쏠림 방지 -> 무작위
np.random.shuffle(data)

4. X, y 데이터로 저장

X= [] # 이미지
y= [] # 라벨(마스크 여부)

for image in data:
  X.append(image[0])
  y.append(image[1])

X = np.array(X)
y = np.array(y)

X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=13)

5. 모델

• LeNet

from tensorflow.keras import layers, models

model = models.Sequential([
    layers.Conv2D(32, kernel_size=(5, 5), strides=(1, 1), padding='same', activation='relu', input_shape=(150, 150, 1)),
    layers.MaxPooling2D(pool_size=(2,2), strides=(2,2)),
    layers.Conv2D(64, (2,2), padding='same', activation='relu'),
    layers.MaxPooling2D(pool_size=(2,2)),
    layers.Dropout(0.25),
    layers.Flatten(),
    layers.Dense(1000, activation='relu'),
    layers.Dense(1, activation='sigmoid')
])

6. compile & fit

from sklearn import metrics
model.compile(
    optimizer='adam', loss=tf.keras.losses.BinaryCrossentropy(), metrics=['accuracy']
)

X_train = X_train.reshape(len(X_train), X_train.shape[1], X_train.shape[2], 1)
X_val = X_val.reshape(len(X_val), X_val.shape[1], X_val.shape[2], 1)

history = model.fit(X_train, y_train, epochs=4, batch_size=32)

model.evaluate(X_val, y_val)
'''
63/63 [==============================] - 1s 9ms/step - loss: 0.0880 - accuracy: 0.9740
[0.08797423541545868, 0.9739999771118164]
'''

7. 결과

prediction = (model.predict(X_val) > 0.5).astype('int32')
print(classification_report(y_val, prediction))
print(confusion_matrix(y_val, prediction))
'''
63/63 [==============================] - 0s 6ms/step
              precision    recall  f1-score   support

           0       0.99      0.96      0.97       998
           1       0.96      0.99      0.97      1002

    accuracy                           0.97      2000
   macro avg       0.97      0.97      0.97      2000
weighted avg       0.97      0.97      0.97      2000

[[955  43]
 [  9 993]]
 '''

8. wrong result

wrong_result = []

for n in range(0, len(y_val)):
  if prediction[n] != y_val[n]:
    wrong_result.append(n)
    
import random

samples = random.sample(wrong_result, 6)

# 이미지 출력
plt.figure(figsize=(14, 12))

for idx, n in enumerate(samples):
  plt.subplot(3, 2, idx+1)
  plt.imshow(X_val[n].reshape(150, 150), interpolation='nearest')
  plt.title(prediction[n])
  plt.axis('off')

plt.show()

---

Reference
1) 제로베이스 데이터스쿨 강의자료