10.4 - Data Augmentation Demos.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Augmentation\n",
    "\n",
    "### Let's look at our untouched dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot images\n",
    "from keras.datasets import mnist\n",
    "from matplotlib import pyplot\n",
    "\n",
    "# load data\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# create a grid of 3x3 images\n",
    "for i in range(0, 9):\n",
    "    pyplot.subplot(330 + 1 + i)\n",
    "    pyplot.imshow(x_train[i], cmap=pyplot.get_cmap('gray'))\n",
    "    \n",
    "# show the plot\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Rotations\n",
    "- As per Keras documentation random is 50%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from matplotlib import pyplot\n",
    "from keras import backend as K\n",
    "\n",
    "# Load data\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape our data to be in the forma [samples, width, height, color_depth]\n",
    "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
    "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
    "\n",
    "# Change datatype to float32\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "\n",
    "# Create our image generator\n",
    "# Define random rotation parameter to be 60 degrees\n",
    "train_datagen = ImageDataGenerator(rotation_range=60)\n",
    "\n",
    "# fit parameters from data\n",
    "train_datagen.fit(x_train)\n",
    "\n",
    "# configure batch size and retrieve one batch of images\n",
    "for x_batch, y_batch in train_datagen.flow(x_train, y_train, batch_size=9):\n",
    "    # create a grid of 3x3 images\n",
    "    for i in range(0, 9):\n",
    "        pyplot.subplot(330 + 1 + i)\n",
    "        pyplot.imshow(x_batch[i].reshape(28, 28), cmap=pyplot.get_cmap('gray'))# show the plot\n",
    "    pyplot.show()\n",
    "    break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Randon Shearing and zooming"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from matplotlib import pyplot\n",
    "from keras import backend as K\n",
    "\n",
    "# Load data\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape our data to be in the forma [samples, width, height, color_depth]\n",
    "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
    "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
    "\n",
    "# Change datatype to float32\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "\n",
    "# Create our image generator\n",
    "# Define shearing and zooming parameters to be 0.5 each\n",
    "train_datagen = ImageDataGenerator(shear_range=0.5,\n",
    "                                   zoom_range=0.5)\n",
    "\n",
    "# fit parameters from data\n",
    "train_datagen.fit(x_train)\n",
    "\n",
    "# configure batch size and retrieve one batch of images\n",
    "for x_batch, y_batch in train_datagen.flow(x_train, y_train, batch_size=9):\n",
    "    # create a grid of 3x3 images\n",
    "    for i in range(0, 9):\n",
    "        pyplot.subplot(330 + 1 + i)\n",
    "        pyplot.imshow(x_batch[i].reshape(28, 28), cmap=pyplot.get_cmap('gray'))# show the plot\n",
    "    pyplot.show()\n",
    "    break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Horizontal and Vertical Flips"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from matplotlib import pyplot\n",
    "from keras import backend as K\n",
    "\n",
    "# Load data\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape our data to be in the forma [samples, width, height, color_depth]\n",
    "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
    "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
    "\n",
    "# Change datatype to float32\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "\n",
    "# define data preparation\n",
    "train_datagen = ImageDataGenerator(vertical_flip=True,\n",
    "                                   horizontal_flip=True)\n",
    "\n",
    "# fit parameters from data\n",
    "train_datagen.fit(x_train)\n",
    "\n",
    "# configure batch size and retrieve one batch of images\n",
    "for x_batch, y_batch in train_datagen.flow(x_train, y_train, batch_size=9):\n",
    "    # create a grid of 3x3 images\n",
    "    for i in range(0, 9):\n",
    "        pyplot.subplot(330 + 1 + i)\n",
    "        pyplot.imshow(x_batch[i].reshape(28, 28), cmap=pyplot.get_cmap('gray'))# show the plot\n",
    "    pyplot.show()\n",
    "    break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Shifts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from matplotlib import pyplot\n",
    "from keras import backend as K\n",
    "\n",
    "# Load data\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape our data to be in the forma [samples, width, height, color_depth]\n",
    "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
    "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
    "\n",
    "# Change datatype to float32\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "\n",
    "# define data preparation\n",
    "train_datagen = ImageDataGenerator(width_shift_range=0.3,\n",
    "                                   height_shift_range=0.3)\n",
    "\n",
    "# fit parameters from data\n",
    "train_datagen.fit(x_train)\n",
    "\n",
    "# configure batch size and retrieve one batch of images\n",
    "for x_batch, y_batch in train_datagen.flow(x_train, y_train, batch_size=9):\n",
    "    # create a grid of 3x3 images\n",
    "    for i in range(0, 9):\n",
    "        pyplot.subplot(330 + 1 + i)\n",
    "        pyplot.imshow(x_batch[i].reshape(28, 28), cmap=pyplot.get_cmap('gray'))# show the plot\n",
    "    pyplot.show()\n",
    "    break"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Applying all at once"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from matplotlib import pyplot\n",
    "from keras import backend as K\n",
    "\n",
    "# Load data\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape our data to be in the forma [samples, width, height, color_depth]\n",
    "x_train = x_train.reshape(x_train.shape[0], 28, 28, 1)\n",
    "x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)\n",
    "\n",
    "# Change datatype to float32\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "\n",
    "# define data preparation\n",
    "train_datagen = ImageDataGenerator(\n",
    "        rotation_range=45,\n",
    "        width_shift_range=0.2,\n",
    "        height_shift_range=0.2,\n",
    "        shear_range=0.2,\n",
    "        zoom_range=0.2,\n",
    "        horizontal_flip=True,\n",
    "        fill_mode='nearest')\n",
    "\n",
    "# fit parameters from data\n",
    "train_datagen.fit(x_train)\n",
    "\n",
    "# configure batch size and retrieve one batch of images\n",
    "for x_batch, y_batch in train_datagen.flow(x_train, y_train, batch_size=9):\n",
    "    # create a grid of 3x3 images\n",
    "    for i in range(0, 9):\n",
    "        pyplot.subplot(330 + 1 + i)\n",
    "        pyplot.imshow(x_batch[i].reshape(28, 28), cmap=pyplot.get_cmap('gray'))# show the plot\n",
    "    pyplot.show()\n",
    "    break\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read more about it at the official Keras Documentation\n",
    "https://keras.io/preprocessing/image/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test Augmentation on a single image\n",
    "- Outputs to ./preview diretory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator, img_to_array, load_img\n",
    "\n",
    "datagen = ImageDataGenerator(\n",
    "        rotation_range=40,\n",
    "        width_shift_range=0.2,\n",
    "        height_shift_range=0.2,\n",
    "        shear_range=0.2,\n",
    "        zoom_range=0.2,\n",
    "        horizontal_flip=True,\n",
    "        fill_mode='nearest')\n",
    "\n",
    "img = load_img('dog.jpeg')  \n",
    "x = img_to_array(img)  # creating a Numpy array with shape (3, 150, 150)\n",
    "x = x.reshape((1,) + x.shape)  # converting to a Numpy array with shape (1, 3, 150, 150)\n",
    "\n",
    "i = 0\n",
    "for batch in datagen.flow(x, save_to_dir='output', save_prefix='dog', save_format='jpeg'):\n",
    "    i += 1\n",
    "    if i > 35:\n",
    "        break "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}