notebooks/resizing-observation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Observation resizing\n",
    "\n",
    "The major mistake in Agent One - a different process of resizing (namely the use of `Image.BILINEAR` :)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Track generation: 1097..1383 -> 286-tiles track\n"
     ]
    }
   ],
   "source": [
    "from gym.envs.box2d.car_racing import CarRacing\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from PIL import Image\n",
    "from scipy.misc import imresize as resize\n",
    "\n",
    "env = CarRacing()\n",
    "o = env.reset()\n",
    "for _ in range(100):\n",
    "    o, r, no, d = env.step(env.action_space.sample())\n",
    "    env.viewer.window.dispatch_events()\n",
    "env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12fa177f0>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "f, a = plt.subplots()\n",
    "proc = ((o).round()).astype(np.uint8)\n",
    "img = Image.fromarray(o).resize((64*3, 64*3))\n",
    "a.imshow(img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12e1c9a20>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "f, a = plt.subplots()\n",
    "a.imshow(o)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Comparing image processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12e1c97b8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "f, a = plt.subplots()\n",
    "a.imshow(o[:, :, :])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Resizing incorrectly:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12e50e710>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "def process_frame(frame, screen_size=None, vertical_cut=-1, max_val=255, save_img=False):\n",
    "    \"\"\" could be two functions  - resizing and saving\"\"\"\n",
    "    frame = frame[:vertical_cut, :, :]\n",
    "    frame = Image.fromarray(frame, mode='RGB')\n",
    "\n",
    "    obs = frame.resize(screen_size)\n",
    "    if save_img:\n",
    "        frame.save(results_dir+'/debug/raw_frame{}.png'.format(save_img), 'PNG')\n",
    "        obs.save(results_dir+'/debug/processed().png'.format(save_img), 'PNG')\n",
    "    return np.array(obs) / max_val\n",
    "\n",
    "f, a = plt.subplots()\n",
    "a.imshow(process_frame(o, screen_size=(64, 64), vertical_cut=84))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12ee40da0>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "def process_frame(frame, screen_size=None, vertical_cut=-1, max_val=255, save_img=False):\n",
    "    \"\"\" could be two functions  - resizing and saving\"\"\"\n",
    "    frame = frame[:vertical_cut, :, :]\n",
    "    frame = Image.fromarray(frame, mode='RGB')\n",
    "\n",
    "    obs = frame.resize(screen_size, Image.BILINEAR)\n",
    "    if save_img:\n",
    "        frame.save(results_dir+'/debug/raw_frame{}.png'.format(save_img), 'PNG')\n",
    "        obs.save(results_dir+'/debug/processed().png'.format(save_img), 'PNG')\n",
    "    return np.array(obs) / max_val\n",
    "\n",
    "f, a = plt.subplots()\n",
    "a.imshow(process_frame(o, screen_size=(64, 64), vertical_cut=84))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/adam/.pyenv/versions/3.6.8/envs/world-models/lib/python3.6/site-packages/ipykernel_launcher.py:3: DeprecationWarning: `imresize` is deprecated!\n",
      "`imresize` is deprecated in SciPy 1.0.0, and will be removed in 1.3.0.\n",
      "Use Pillow instead: ``numpy.array(Image.fromarray(arr).resize())``.\n",
      "  This is separate from the ipykernel package so we can avoid doing imports until\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x131b2a5f8>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "def _process_frame(frame):\n",
    "    obs = frame[0:84, :, :].astype(np.float)/255.0\n",
    "    obs = resize(obs, (64, 64))\n",
    "    obs = ((1.0 - obs) * 255).round().astype(np.uint8)\n",
    "    return obs\n",
    "\n",
    "f, a = plt.subplots()\n",
    "a.imshow(_process_frame(o))"
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}