numpy_note-01-Padding-function-in-numpy.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# The padding function in Numpy\n",
    "- We want to show this function how to be utilized on the padding operations of convolutional neural networks.\n",
    "\n",
    "## No channel\n",
    "- Example for no channel with 2-dimensional matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(Original) 2d feature map=>\n",
      "[[1 2 3]\n",
      " [4 5 6]\n",
      " [7 8 9]]\n",
      "\n",
      "(Padding)  2d feature map=>\n",
      "[[0 0 0 0 0]\n",
      " [0 1 2 3 0]\n",
      " [0 4 5 6 0]\n",
      " [0 7 8 9 0]\n",
      " [0 0 0 0 0]]\n"
     ]
    }
   ],
   "source": [
    "a_2d = np.arange(1,10).reshape(3,3)\n",
    "print(\"(Original) 2d feature map=>\\n{0}\\n\".format(a_2d))\n",
    "\n",
    "a_2d = np.pad(array=a_2d, pad_width=((1,1), (1,1)), mode='constant', constant_values=0)\n",
    "print(\"(Padding)  2d feature map=>\\n{0}\".format(a_2d))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Channel first\n",
    "- Example for a 1-channel feature map with channel first setting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(Original) 3d feature map=>\n",
      "[[[1 2 3]\n",
      "  [4 5 6]\n",
      "  [7 8 9]]]\n",
      "\n",
      "(Padding)  3d feature map=>\n",
      "[[[0 0 0 0 0]\n",
      "  [0 1 2 3 0]\n",
      "  [0 4 5 6 0]\n",
      "  [0 7 8 9 0]\n",
      "  [0 0 0 0 0]]]\n"
     ]
    }
   ],
   "source": [
    "a_3d = np.arange(1,10).reshape(1,3,3)\n",
    "print(\"(Original) 3d feature map=>\\n{0}\\n\".format(a_3d))\n",
    "\n",
    "a_3d = np.pad(array=a_3d, pad_width=((0,0),(1,1),(1,1)), mode='constant', constant_values=0)\n",
    "print(\"(Padding)  3d feature map=>\\n{0}\".format(a_3d))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Example for a 2-channel feature map with **channel first** setting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(Original) 3d feature map=>\n",
      "[[[ 1  2  3]\n",
      "  [ 4  5  6]\n",
      "  [ 7  8  9]]\n",
      "\n",
      " [[10 11 12]\n",
      "  [13 14 15]\n",
      "  [16 17 18]]]\n",
      "\n",
      "(Padding)  3d feature map=>\n",
      "[[[ 0  0  0  0  0]\n",
      "  [ 0  1  2  3  0]\n",
      "  [ 0  4  5  6  0]\n",
      "  [ 0  7  8  9  0]\n",
      "  [ 0  0  0  0  0]]\n",
      "\n",
      " [[ 0  0  0  0  0]\n",
      "  [ 0 10 11 12  0]\n",
      "  [ 0 13 14 15  0]\n",
      "  [ 0 16 17 18  0]\n",
      "  [ 0  0  0  0  0]]]\n"
     ]
    }
   ],
   "source": [
    "a_3d = np.arange(1,19).reshape(2,3,3)\n",
    "print(\"(Original) 3d feature map=>\\n{0}\\n\".format(a_3d))\n",
    "\n",
    "a_3d = np.pad(array=a_3d, pad_width=((0,0),(1,1),(1,1)), mode='constant', constant_values=0)\n",
    "print(\"(Padding)  3d feature map=>\\n{0}\".format(a_3d))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Example for a 3-channel feature map with **channel first** setting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(Original) 3d feature map=>\n",
      "[[[ 1  2  3]\n",
      "  [ 4  5  6]\n",
      "  [ 7  8  9]]\n",
      "\n",
      " [[10 11 12]\n",
      "  [13 14 15]\n",
      "  [16 17 18]]\n",
      "\n",
      " [[19 20 21]\n",
      "  [22 23 24]\n",
      "  [25 26 27]]]\n",
      "\n",
      "(Padding)  3d feature map=>\n",
      "[[[ 0  0  0  0  0]\n",
      "  [ 0  1  2  3  0]\n",
      "  [ 0  4  5  6  0]\n",
      "  [ 0  7  8  9  0]\n",
      "  [ 0  0  0  0  0]]\n",
      "\n",
      " [[ 0  0  0  0  0]\n",
      "  [ 0 10 11 12  0]\n",
      "  [ 0 13 14 15  0]\n",
      "  [ 0 16 17 18  0]\n",
      "  [ 0  0  0  0  0]]\n",
      "\n",
      " [[ 0  0  0  0  0]\n",
      "  [ 0 19 20 21  0]\n",
      "  [ 0 22 23 24  0]\n",
      "  [ 0 25 26 27  0]\n",
      "  [ 0  0  0  0  0]]]\n"
     ]
    }
   ],
   "source": [
    "a_3d = np.arange(1,28).reshape(3,3,3)\n",
    "print(\"(Original) 3d feature map=>\\n{0}\\n\".format(a_3d))\n",
    "\n",
    "a_3d = np.pad(array=a_3d, pad_width=((0,0),(1,1),(1,1)), mode='constant', constant_values=0)\n",
    "print(\"(Padding)  3d feature map=>\\n{0}\".format(a_3d))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Example for two 3-channel feature maps with **channel first** setting. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(Original) 3d feature map=>\n",
      "[[[[ 1  2  3]\n",
      "   [ 4  5  6]\n",
      "   [ 7  8  9]]\n",
      "\n",
      "  [[10 11 12]\n",
      "   [13 14 15]\n",
      "   [16 17 18]]\n",
      "\n",
      "  [[19 20 21]\n",
      "   [22 23 24]\n",
      "   [25 26 27]]]\n",
      "\n",
      "\n",
      " [[[28 29 30]\n",
      "   [31 32 33]\n",
      "   [34 35 36]]\n",
      "\n",
      "  [[37 38 39]\n",
      "   [40 41 42]\n",
      "   [43 44 45]]\n",
      "\n",
      "  [[46 47 48]\n",
      "   [49 50 51]\n",
      "   [52 53 54]]]]\n",
      "\n",
      "(Padding)  3d feature map=>\n",
      "[[[[ 0  0  0  0  0]\n",
      "   [ 0  1  2  3  0]\n",
      "   [ 0  4  5  6  0]\n",
      "   [ 0  7  8  9  0]\n",
      "   [ 0  0  0  0  0]]\n",
      "\n",
      "  [[ 0  0  0  0  0]\n",
      "   [ 0 10 11 12  0]\n",
      "   [ 0 13 14 15  0]\n",
      "   [ 0 16 17 18  0]\n",
      "   [ 0  0  0  0  0]]\n",
      "\n",
      "  [[ 0  0  0  0  0]\n",
      "   [ 0 19 20 21  0]\n",
      "   [ 0 22 23 24  0]\n",
      "   [ 0 25 26 27  0]\n",
      "   [ 0  0  0  0  0]]]\n",
      "\n",
      "\n",
      " [[[ 0  0  0  0  0]\n",
      "   [ 0 28 29 30  0]\n",
      "   [ 0 31 32 33  0]\n",
      "   [ 0 34 35 36  0]\n",
      "   [ 0  0  0  0  0]]\n",
      "\n",
      "  [[ 0  0  0  0  0]\n",
      "   [ 0 37 38 39  0]\n",
      "   [ 0 40 41 42  0]\n",
      "   [ 0 43 44 45  0]\n",
      "   [ 0  0  0  0  0]]\n",
      "\n",
      "  [[ 0  0  0  0  0]\n",
      "   [ 0 46 47 48  0]\n",
      "   [ 0 49 50 51  0]\n",
      "   [ 0 52 53 54  0]\n",
      "   [ 0  0  0  0  0]]]]\n"
     ]
    }
   ],
   "source": [
    "multi_a_3d = np.arange(1,55).reshape(2,3,3,3)\n",
    "print(\"(Original) 3d feature map=>\\n{0}\\n\".format(multi_a_3d))\n",
    "\n",
    "multi_a_3d = np.pad(array=multi_a_3d, pad_width=((0,0),(0,0),(1,1),(1,1)), mode='constant', constant_values=0)\n",
    "print(\"(Padding)  3d feature map=>\\n{0}\".format(multi_a_3d))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Channel last\n",
    "- Example for two 3-channel feature maps with **channel last** setting.\n",
    "- This example quite like the real situation, because of generally the ordering of image's channel are set with channel-last."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(Original) 3d feature map=> Shape:(1, 3, 3, 3)\n",
      "[[[[0.2 0.2 0.2]\n",
      "   [1.  1.  1. ]\n",
      "   [0.2 0.2 0.2]]\n",
      "\n",
      "  [[0.5 0.5 0.5]\n",
      "   [1.  1.  1. ]\n",
      "   [0.5 0.5 0.5]]\n",
      "\n",
      "  [[0.8 0.8 0.8]\n",
      "   [1.  1.  1. ]\n",
      "   [0.8 0.8 0.8]]]]\n",
      "(Padding)  3d feature map=> Shape:(1, 5, 5, 3)\n",
      "[[[[0.  0.  0. ]\n",
      "   [0.  0.  0. ]\n",
      "   [0.  0.  0. ]\n",
      "   [0.  0.  0. ]\n",
      "   [0.  0.  0. ]]\n",
      "\n",
      "  [[0.  0.  0. ]\n",
      "   [0.2 0.2 0.2]\n",
      "   [1.  1.  1. ]\n",
      "   [0.2 0.2 0.2]\n",
      "   [0.  0.  0. ]]\n",
      "\n",
      "  [[0.  0.  0. ]\n",
      "   [0.5 0.5 0.5]\n",
      "   [1.  1.  1. ]\n",
      "   [0.5 0.5 0.5]\n",
      "   [0.  0.  0. ]]\n",
      "\n",
      "  [[0.  0.  0. ]\n",
      "   [0.8 0.8 0.8]\n",
      "   [1.  1.  1. ]\n",
      "   [0.8 0.8 0.8]\n",
      "   [0.  0.  0. ]]\n",
      "\n",
      "  [[0.  0.  0. ]\n",
      "   [0.  0.  0. ]\n",
      "   [0.  0.  0. ]\n",
      "   [0.  0.  0. ]\n",
      "   [0.  0.  0. ]]]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x11918c588>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "a_3d = np.array([[ [[0.2,0.2,0.2],[  1,  1,  1],[0.2,0.2,0.2]],\n",
    "                   [[0.5,0.5,0.5],[  1,  1,  1],[0.5,0.5,0.5]],\n",
    "                   [[0.8,0.8,0.8],[  1,  1,  1],[0.8,0.8,0.8]]]])\n",
    "print(\"(Original) 3d feature map=> Shape:{0}\".format(a_3d.shape))\n",
    "print(a_3d)\n",
    "\n",
    "fig, axarr = plt.subplots(1, 2)\n",
    "axarr[0].set_title('a_3d')\n",
    "axarr[0].imshow(a_3d[0,:,:,:])\n",
    "\n",
    "a_3d = np.pad(array=a_3d, pad_width=((0,0),(1,1),(1,1),(0,0)), mode='constant', constant_values=0)\n",
    "print(\"(Padding)  3d feature map=> Shape:{0}\".format(a_3d.shape))\n",
    "print(a_3d)\n",
    "\n",
    "axarr[1].set_title('Padded a_3d')\n",
    "axarr[1].imshow(a_3d[0,:,:,:])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Coursera",
   "language": "python",
   "name": "coursera"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}