PDF-Noise-Classification

To decrease the time taken by a manual pre-OCR pdf labeling process. I built a CNN image classification model using OpenCV, Tensorflow and Keras in Python using Jyputer Notebooks. The CNN model classifies PDFs according to the amount of noise, 1 - representing very high noise and 5 - representing clear pdfs.

loading PDFS

images = []
labels = []
image_folder_Path =""
for image_file_path in imutils.paths.list_images(image_folder_Path):
    image_file = cv2.imread(image_file_path, cv2.IMREAD_GRAYSCALE)
    resized_img = cv2.resize(image_file,(760,1000))
    label = image_file_path.split('/')[9]
    images.append(resized_img)
    labels.append(label)

Fixing PDF Labels

pdf_labels_dict = {
    'Q01_unreadable_text': 0,
    'Q02_unclear_text_breaky_sticky': 1,
    'Q03_semi_clear_text': 2,
    'Q04_clear_text': 3,
    'Q05_perfectly_clear_text': 4,
}
numbered_Labels = []
for label in labels:
    label = pdf_labels_dict[label]
    numbered_Labels.append(label)
labels =numbered_Labels

PDF examples

Train Test Split

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)

Model Building

num_classes = 5
input_shape=(5, 1000, 760,1)
model = Sequential([
  layers.Conv2D(16, 3, padding='same', activation='relu', input_shape=input_shape[1:]),
  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'])

Fit Training Set

model.fit(X_train_scaled, y_train, epochs=30) 

|Epoch|Accuracy| | 1 | 0.2634 | | 7 | 0.7737 | | 11 | 0.9918 | | 15 | 1.0000 | | 30 | 1.0000 |

Fit Test Set

model.evaluate(X_test_scaled,y_test)

accuracy: 0.3537

Add Data Augmentation

data_augmentation = keras.Sequential(
  [
    layers.experimental.preprocessing.RandomFlip("horizontal", 
                                                 input_shape=(1000, 
                                                              760,
                                                              1)),
    layers.experimental.preprocessing.RandomRotation(0.1),
    layers.experimental.preprocessing.RandomZoom(0.1),
  ]
)

Rebuild Model with Data Augmentation

num_classes = 5
input_shape=(5, 1000, 760,1)

model2 = Sequential([
  data_augmentation,

  #Convolutional Network
  layers.Conv2D(16, 3, padding='same', activation='relu', input_shape=input_shape[1:]),
  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.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(),

  #Dense network
  layers.Flatten(),
  layers.Dense(5000, activation='sigmoid'),
  layers.Dense(500, activation='sigmoid'),
  layers.Dense(50, activation='sigmoid'),
  layers.Dense(num_classes,activation='softmax' )
])

model2.compile(optimizer='SGD',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

ReFit Training Set

model2.fit(X_train_scaled, y_train, epochs=500) 

|Epoch|Accuracy| | 1 | 0.1152 | | 2 | 0.2510| | 3 | 0.3292 | | 300 | 0.3292 | | 500 | 0.3292 |

Refit Test Set

model2.evaluate(X_test_scaled,y_test)

accuracy: 0.4268