ImageDataGenerator_Example.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ImageDataGenerator_Example.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jvishnuvardhan/Keras_Examples/blob/master/ImageDataGenerator_Example.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "chfH4o5Zbbc2",
        "colab_type": "code",
        "outputId": "a5d8a65e-9b19-421e-b914-6c0c133c5fb7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 380
        }
      },
      "source": [
        "!pip install tensorflow==2.0.0rc1"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: tensorflow==2.0.0rc1 in /usr/local/lib/python3.6/dist-packages (2.0.0rc1)\n",
            "Requirement already satisfied: google-pasta>=0.1.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (0.1.7)\n",
            "Requirement already satisfied: protobuf>=3.6.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (3.7.1)\n",
            "Requirement already satisfied: absl-py>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (0.8.0)\n",
            "Requirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.1.0)\n",
            "Requirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.15.0)\n",
            "Requirement already satisfied: wrapt>=1.11.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.11.2)\n",
            "Requirement already satisfied: tf-estimator-nightly<1.14.0.dev2019080602,>=1.14.0.dev2019080601 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.14.0.dev2019080601)\n",
            "Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (0.33.6)\n",
            "Requirement already satisfied: tb-nightly<1.15.0a20190807,>=1.15.0a20190806 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.15.0a20190806)\n",
            "Requirement already satisfied: gast>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (0.2.2)\n",
            "Requirement already satisfied: keras-applications>=1.0.8 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.0.8)\n",
            "Requirement already satisfied: numpy<2.0,>=1.16.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.16.5)\n",
            "Requirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.12.0)\n",
            "Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (1.1.0)\n",
            "Requirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (0.8.0)\n",
            "Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.6/dist-packages (from tensorflow==2.0.0rc1) (3.0.1)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from protobuf>=3.6.1->tensorflow==2.0.0rc1) (41.2.0)\n",
            "Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.15.0a20190807,>=1.15.0a20190806->tensorflow==2.0.0rc1) (3.1.1)\n",
            "Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.15.0a20190807,>=1.15.0a20190806->tensorflow==2.0.0rc1) (0.15.6)\n",
            "Requirement already satisfied: h5py in /usr/local/lib/python3.6/dist-packages (from keras-applications>=1.0.8->tensorflow==2.0.0rc1) (2.8.0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P7wdnwL9bcIU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "mnist = tf.keras.datasets.mnist\n",
        "\n",
        "(x_train, y_train),(x_test, y_test) = mnist.load_data()\n",
        "x_train, x_test = x_train / 255.0, x_test / 255.0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TIXbsPs-bgQa",
        "colab_type": "code",
        "outputId": "c8879aa6-0ba4-48ee-c578-d67325e81689",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "# create a grid of 3x3 images\n",
        "for i in range(0,9):\n",
        "  plt.subplot(330+1+i)\n",
        "  plt.imshow(x_train[i], cmap=plt.get_cmap('gray'))\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqIhIBKpERUQiUCUqIhLB/wOA2s4TjzPBwgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WIUhuA2YcQSl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras_preprocessing.image import ImageDataGenerator\n",
        "from keras import backend as K"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nkTkVxVFdmYp",
        "colab_type": "code",
        "outputId": "cd108833-388c-4863-84c7-f7c3f498ba47",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 358
        }
      },
      "source": [
        "#define data preparation\n",
        "datagen=ImageDataGenerator(featurewise_center=True,featurewise_std_normalization=True)\n",
        "(x_train, y_train),(x_test, y_test) = mnist.load_data()\n",
        "x_train, x_test = x_train / 255.0, x_test / 255.0\n",
        "x_train=x_train.reshape(x_train.shape[0],1,28,28)\n",
        "#fit parameters\n",
        "datagen.fit(x_train,seed=10)\n",
        "# configure and retrieve batches of images\n",
        "for x_batch, y_batch in datagen.flow(x_train,y_train,batch_size=9,shuffle=False):\n",
        "  for i in range(0,9):\n",
        "    plt.subplot(330+1+i)\n",
        "    plt.imshow(x_batch[i].reshape(28,28), cmap=plt.get_cmap('gray'))\n",
        "  plt.show()\n",
        "  break"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/image_data_generator.py:940: UserWarning: Expected input to be images (as Numpy array) following the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3 or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  ' channels).')\n",
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/numpy_array_iterator.py:127: UserWarning: NumpyArrayIterator is set to use the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3, or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  str(self.x.shape[channels_axis]) + ' channels).')\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nv7b8AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I4gIgzUjfwb-",
        "colab_type": "code",
        "outputId": "d700918d-9a12-40ac-bb6e-398b48fc1f12",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        }
      },
      "source": [
        "#ZCA whitening\n",
        "(x_train, y_train),(x_test, y_test) = mnist.load_data()\n",
        "x_train, x_test = x_train / 255.0, x_test / 255.0\n",
        "x_train=x_train.reshape(x_train.shape[0],1,28,28)\n",
        "datagen = ImageDataGenerator(zca_whitening=True)\n",
        "datagen.fit(x_train)\n",
        "# configure batch size and retrieve one batch of images\n",
        "for X_batch, y_batch in datagen.flow(x_train, y_train, batch_size=9,shuffle=False):\n",
        "# create a grid of 3x3 images\n",
        "  for i in range(0, 9):\n",
        "    plt.subplot(330 + 1 + i)\n",
        "    plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))\n",
        "    # show the plot\n",
        "  plt.show()\n",
        "  break"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/image_data_generator.py:336: UserWarning: This ImageDataGenerator specifies `zca_whitening`, which overrides setting of `featurewise_center`.\n",
            "  warnings.warn('This ImageDataGenerator specifies '\n",
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/image_data_generator.py:940: UserWarning: Expected input to be images (as Numpy array) following the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3 or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  ' channels).')\n",
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/numpy_array_iterator.py:127: UserWarning: NumpyArrayIterator is set to use the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3, or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  str(self.x.shape[channels_axis]) + ' channels).')\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\niDqO47SBP0Qdx3Ha4P8HCBH0vkjrBUUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1ULf3133glfc",
        "colab_type": "code",
        "outputId": "190f39cb-2ebd-4229-ffc8-bff766d0bf3d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 358
        }
      },
      "source": [
        "#Random rotations\n",
        "datagen = ImageDataGenerator(rotation_range=90)\n",
        "(x_train, y_train),(x_test, y_test) = mnist.load_data()\n",
        "x_train, x_test = x_train / 255.0, x_test / 255.0\n",
        "x_train=x_train.reshape(x_train.shape[0],1,28,28)\n",
        "datagen.fit(x_train)\n",
        "# configure batch size and retrieve one batch of images\n",
        "for X_batch, y_batch in datagen.flow(x_train, y_train, batch_size=9,shuffle=False):\n",
        "# create a grid of 3x3 images\n",
        "  for i in range(0, 9):\n",
        "    plt.subplot(330 + 1 + i)\n",
        "    plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))\n",
        "    # show the plot\n",
        "  plt.show()\n",
        "  break"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/image_data_generator.py:940: UserWarning: Expected input to be images (as Numpy array) following the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3 or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  ' channels).')\n",
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/numpy_array_iterator.py:127: UserWarning: NumpyArrayIterator is set to use the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3, or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  str(self.x.shape[channels_axis]) + ' channels).')\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRC06eP1AAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D1_fOsYLgvXR",
        "colab_type": "code",
        "outputId": "d58e96c1-5a8a-4be2-8c73-194a6289caa3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 358
        }
      },
      "source": [
        "#Random shifts\n",
        "shift = 0.2\n",
        "datagen = ImageDataGenerator(width_shift_range=shift, height_shift_range=shift)\n",
        "(x_train, y_train),(x_test, y_test) = mnist.load_data()\n",
        "x_train, x_test = x_train / 255.0, x_test / 255.0\n",
        "x_train=x_train.reshape(x_train.shape[0],1,28,28)\n",
        "datagen.fit(x_train)\n",
        "# configure batch size and retrieve one batch of images\n",
        "for X_batch, y_batch in datagen.flow(x_train, y_train, batch_size=9,shuffle=False):\n",
        "# create a grid of 3x3 images\n",
        "  for i in range(0, 9):\n",
        "    plt.subplot(330 + 1 + i)\n",
        "    plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))\n",
        "    # show the plot\n",
        "  plt.show()\n",
        "  break"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/image_data_generator.py:940: UserWarning: Expected input to be images (as Numpy array) following the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3 or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  ' channels).')\n",
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/numpy_array_iterator.py:127: UserWarning: NumpyArrayIterator is set to use the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3, or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  str(self.x.shape[channels_axis]) + ' channels).')\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RULjnpKy-wjQ",
        "colab_type": "code",
        "outputId": "f2850658-9d23-4043-e512-23372ba26384",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 358
        }
      },
      "source": [
        "#Random flips\n",
        "shift = 0.2\n",
        "datagen = ImageDataGenerator(horizontal_flip=True, vertical_flip=True)\n",
        "(x_train, y_train),(x_test, y_test) = mnist.load_data()\n",
        "x_train, x_test = x_train / 255.0, x_test / 255.0\n",
        "x_train=x_train.reshape(x_train.shape[0],1,28,28)\n",
        "datagen.fit(x_train)\n",
        "# configure batch size and retrieve one batch of images\n",
        "for X_batch, y_batch in datagen.flow(x_train, y_train, batch_size=9,shuffle=False):\n",
        "# create a grid of 3x3 images\n",
        "  for i in range(0, 9):\n",
        "    plt.subplot(330 + 1 + i)\n",
        "    plt.imshow(X_batch[i].reshape(28, 28), cmap=plt.get_cmap('gray'))\n",
        "    # show the plot\n",
        "  plt.show()\n",
        "  break"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/image_data_generator.py:940: UserWarning: Expected input to be images (as Numpy array) following the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3 or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  ' channels).')\n",
            "/usr/local/lib/python3.6/dist-packages/keras_preprocessing/image/numpy_array_iterator.py:127: UserWarning: NumpyArrayIterator is set to use the data format convention \"channels_last\" (channels on axis 3), i.e. expected either 1, 3, or 4 channels on axis 3. However, it was passed an array with shape (60000, 1, 28, 28) (28 channels).\n",
            "  str(self.x.shape[channels_axis]) + ' channels).')\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEYlBN1ERkRh0ExURieH/AdARi129xTIUAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}