MNIST_data_aug_image_shiftting.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8c95b0b0-1706-4e2c-878f-6ccb54c10db1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['data', 'target', 'frame', 'categories', 'feature_names', 'target_names', 'DESCR', 'details', 'url'])"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.datasets import fetch_openml\n",
    "mnist = fetch_openml('mnist_784', version=1, as_frame=False)\n",
    "mnist.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d4fab03a-d847-4cdd-a78e-8ad4e5bdecfa",
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = mnist[\"data\"], mnist[\"target\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8ab32580-2c64-4f04-98c5-59cca6e8e83f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000, 784)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "57417f5e-7c94-44fc-90a4-ab7434754b8b",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = X[:60000], X[60000:], y[:60000], y[60000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c508de86-8c9e-4900-b44f-b52f88c031a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.ndimage.interpolation import shift"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "623ef909-6482-4f08-b935-53106e28c23e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def shift_image(image, dx, dy):\n",
    "    image = image.reshape((28, 28))\n",
    "    shifted_image = shift(image, [dy, dx], cval=0, mode='constant')\n",
    "    return shifted_image.reshape([-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4e516531-4e4d-4b90-a7db-ccc52c556b14",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "image = X_train[1000]\n",
    "shifted_image_down = shift_image(image, 0, 5)\n",
    "shifted_image_left = shift_image(image, -5, 0)\n",
    "shifted_image_right = shift_image(image, 7, 0)\n",
    "\n",
    "plt.figure(figsize=(12, 3))\n",
    "plt.subplot(141)\n",
    "plt.title(\"original\", fontsize=14)\n",
    "plt.imshow(image.reshape(28, 28), interpolation='nearest', cmap='Greys')\n",
    "plt.subplot(142)\n",
    "plt.title(\"Shifted down\", fontsize=14)\n",
    "plt.imshow(shifted_image_down.reshape(28, 28), interpolation='nearest', cmap='Greys')\n",
    "plt.subplot(143)\n",
    "plt.title(\"Shifted left\", fontsize=14)\n",
    "plt.imshow(shifted_image_left.reshape(28, 28), interpolation='nearest', cmap='Greys')\n",
    "plt.subplot(144)\n",
    "plt.title(\"Shifted right\", fontsize=14)\n",
    "plt.imshow(shifted_image_right.reshape(28, 28), interpolation='nearest', cmap='Greys')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "79e06ea4-fd46-40cd-9482-bbf1e61c6b1b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_16665/1481486108.py:12: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
      "  y_train_augmented = np.array(y_train_augmented)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "X_train_augmented = [image for image in X_train]\n",
    "y_train_augmented = [label for label in X_train]\n",
    "\n",
    "for dx, dy in ((1, 0), (-1, 0), (0, 1), (0, -1)):\n",
    "    for image, label in zip(X_train, y_train):\n",
    "        X_train_augmented.append(shift_image(image, dx, dy))\n",
    "        y_train_augmented.append(label)\n",
    "        \n",
    "X_train_augmented = np.array(X_train_augmented)\n",
    "y_train_augmented = np.array(y_train_augmented)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "c0efc6b6-b646-4fe9-b88c-f9739e54d736",
   "metadata": {},
   "outputs": [],
   "source": [
    "shuffle_idx = np.random.permutation(len(X_train_augmented))\n",
    "X_train_augmented = X_train_augmented[shuffle_idx]\n",
    "y_train_augmented = y_train_augmented[shuffle_idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "e816430c-c30d-48b9-a047-4ccea22d75a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_image(data):\n",
    "    image = data.reshape(28, 28)\n",
    "    plt.imshow(image, cmap=mpl.cm.binary)\n",
    "    \n",
    "    \n",
    "im = X_train_augmented[0]\n",
    "plot_image(im)\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}