error_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "source": [
    "# File for doing error analysis on the gaze estimation\n",
    "# predictions on augmented dataset. \n",
    "# Author : Srinivas Kumar (srinivas1996kumar@gmail.com) \n",
    "\n",
    "import h5py\n",
    "import numpy as np\n",
    "import pickle\n",
    "\n",
    "import matplotlib.pyplot as plt\n"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "source": [
    "# Getting the original gazes and original image count\n",
    "\n",
    "gaze_list = []\n",
    "gaze_original_list = []\n",
    "\n",
    "person_cnt = 0\n",
    "\n",
    "#Change to the appropriate samples once the redirections have been done. \n",
    "aug_file = \"/bigpool/fachpraktikum2021/group5/dir_srinivas/outputs_sted/Redirected_samples1.h5\"\n",
    "total_imgs = 0\n",
    "\n",
    "with h5py.File(aug_file, 'r') as f:\n",
    "\n",
    "    \n",
    "    for idx, (person_id, group) in enumerate(f.items()):\n",
    "        #print(group[\"image\"].shape)\n",
    "\n",
    "        total_imgs += len(group[\"image\"])\n",
    "        \n",
    "        person_cnt += 1\n",
    "\n",
    "\n",
    "        gaze_list.extend( group[\"gaze\"][:])\n",
    "        #gaze_original_list.append(group[\"gaze\"][:])\n",
    "\n",
    "        #print(\"Finished processing \" + str(person_cnt))\n",
    "\n",
    "print(\"The total number of images are: \" + str(total_imgs))\n",
    "print(\"The person count is: \" + str(person_cnt))\n",
    "\n",
    "\n",
    "total_imgs = 0\n",
    "person_cnt = 0\n",
    "\n",
    "with h5py.File(\"/bigpool/fachpraktikum2021/group5/dir_srinivas/outputs_sted/GazeCapture.h5\", 'r') as f:\n",
    "\n",
    "    \n",
    "    for idx, (person_id, group) in enumerate(f.items()):\n",
    "        #print(group[\"image\"].shape)\n",
    "\n",
    "        \n",
    "        person_cnt += 1\n",
    "        total_imgs += len(group[\"pixels\"])\n",
    "\n",
    "\n",
    "        #gaze_list.extend( group[\"gaze\"][:])\n",
    "        gaze_original_list.extend(group[\"labels\"][:, :2])\n",
    "\n",
    "        #print(\"Finished processing \" + str(person_cnt))\n",
    "\n",
    "print(\"The total number of images are: \" + str(total_imgs))\n",
    "print(\"The person count is: \" + str(person_cnt))\n",
    "\n",
    "\n"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "The total number of images are: 2758166\n",
      "The person count is: 1177\n",
      "The total number of images are: 1634443\n",
      "The person count is: 1364\n"
     ]
    }
   ],
   "metadata": {
    "tags": []
   }
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "source": [
    "# Some metrics on number of values outside a range.  \n",
    "# (This cell is for the original dataset not for the augmented dataset)\n",
    "\n",
    "gaze_original_np = np.array(gaze_original_list)\n",
    "\n",
    "# Conversion from radians to degrees \n",
    "gaze_np = gaze_original_np*(180/ np.pi)\n",
    "\n",
    "pitch_original_np = gaze_np[:, 0]\n",
    "yaw_original_np = gaze_np[:, 1]\n",
    "\n",
    "print(\"The len is \" + str(len(pitch_original_np)))\n",
    "print(\"The max pitch is : \" + str(np.max(pitch_original_np)))\n",
    "print(\"The max yaw is: \" + str(np.max(yaw_original_np)))\n",
    "print(\"The min pitch is : \" + str(np.min(pitch_original_np)))\n",
    "print(\"The min yaw is: \" + str(np.min(yaw_original_np)))\n",
    "\n",
    "min_ang = -20\n",
    "max_ang = 20\n",
    "\n",
    "# The number of values outside the range for either pitch/ yaw. \n",
    "lesser = pitch_original_np < min_ang\n",
    "greater = pitch_original_np > max_ang \n",
    "\n",
    "condition1 = np.logical_or(lesser, greater)\n",
    "#print(condition1)\n",
    "\n",
    "lesser = yaw_original_np < min_ang\n",
    "greater = yaw_original_np > max_ang \n",
    "\n",
    "condition2 = np.logical_or(lesser, greater)\n",
    "#print(condition2)\n",
    "\n",
    "condition = np.logical_or(condition1, condition2)\n",
    "\n",
    "print(\"The number of values outside the range(-20, 20) is \"  + str(np.sum(condition)))\n",
    "\n"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "The len is 1634443\n",
      "The max pitch is : 49.524464\n",
      "The max yaw is: 47.646915\n",
      "The min pitch is : -41.620586\n",
      "The min yaw is: -45.41518\n",
      "The number of values outside the range(-20, 20) is 97496\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "source": [
    "# Some metrics on number of values outside a range.  \n",
    "# (This cell is for the augmented dataset)\n",
    "\n",
    "\n",
    "\n",
    "gaze_np = np.array(gaze_list)\n",
    "\n",
    "# Conversion from radians to degrees \n",
    "gaze_np = gaze_np*(180/ np.pi)\n",
    "\n",
    "pitch_np = gaze_np[:, 0]\n",
    "yaw_np = gaze_np[:, 1]\n",
    "\n",
    "print(\"The max pitch is : \" + str(np.max(pitch_np)))\n",
    "print(\"The max yaw is: \" + str(np.max(yaw_np)))\n",
    "print(\"The min pitch is : \" + str(np.min(pitch_np)))\n",
    "print(\"The min yaw is: \" + str(np.min(yaw_np)))\n",
    "\n",
    "\n",
    "min_ang = -20\n",
    "max_ang = 20\n",
    "\n",
    "\n",
    "lesser = pitch_np < min_ang\n",
    "greater = pitch_np > max_ang \n",
    "\n",
    "condition1 = np.logical_or(lesser, greater)\n",
    "#print(condition1)\n",
    "\n",
    "lesser = yaw_np < min_ang\n",
    "greater = yaw_np > max_ang \n",
    "\n",
    "condition2 = np.logical_or(lesser, greater)\n",
    "#print(condition2)\n",
    "\n",
    "condition = np.logical_or(condition1, condition2)\n",
    "\n",
    "print(\"The number of values outside the range(-20, 20) is \"  + str(np.sum(condition)))"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "The max pitch is : 50.07624\n",
      "The max yaw is: 48.275383\n",
      "The min pitch is : -44.063137\n",
      "The min yaw is: -45.41518\n",
      "The number of values outside the range(-20, 20) is 168448\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "source": [
    "# The number of predictions displaying high error. \n",
    "\n",
    "# Load in the prediction losses for the Augmented Dataset\n",
    "with open('/bigpool/fachpraktikum2021/group5/dir_srinivas/outputs_sted/loss_2303196', 'rb') as f:\n",
    "    loss_list = pickle.load(f)\n",
    "\n",
    "loses_np = np.array(loss_list)\n",
    "losses_mean = np.mean(loses_np)\n",
    "losses_std = np.std(loses_np)\n",
    "\n",
    "# Since the state of art is 4.3 degrees \n",
    "baseline_deg = 4.3\n",
    "\n",
    "# Error Percentages for mean + (n)*std\n",
    "error_per_1 = []\n",
    "error_per_2 = []\n",
    "error_per_3 = []\n",
    "\n",
    "# Error percentage for baseline\n",
    "error_per_baseline = []\n",
    "\n",
    "\n",
    "threshold_list = []\n",
    "sample_cnt_list = []\n",
    "\n",
    "for threshold in range(10, 50):\n",
    "    # Vary the threshold from 10 degrees to 50 degrees and for all the values outside this range count \n",
    "    # the number of values displaying high error. \n",
    "\n",
    "    min_ang = -threshold\n",
    "    max_ang = threshold\n",
    "\n",
    "\n",
    "    lesser = pitch_np < min_ang\n",
    "    greater = pitch_np > max_ang \n",
    "\n",
    "    condition1 = np.logical_or(lesser, greater)\n",
    "    #print(condition1)\n",
    "\n",
    "    lesser = yaw_np < min_ang\n",
    "    greater = yaw_np > max_ang \n",
    "\n",
    "    condition2 = np.logical_or(lesser, greater)\n",
    "    #print(condition2)\n",
    "\n",
    "    condition = np.logical_or(condition1, condition2)\n",
    "\n",
    "    #print(\"The number of values outside the range is \"  + str(np.sum(condition)))\n",
    "\n",
    "\n",
    "    # All the samples displaying high error(i.e loss > mean + n*(std))\n",
    "\n",
    "    high_error1 = np.sum(loses_np[condition] > losses_mean + 1*losses_std)\n",
    "    high_error2 = np.sum(loses_np[condition] > losses_mean + 2*losses_std)\n",
    "    high_error3 = np.sum(loses_np[condition] > losses_mean + 3*losses_std)\n",
    "\n",
    "    high_error_baseline = np.sum(loses_np[condition] >  baseline_deg)\n",
    "\n",
    "    sample_cnt = np.sum(condition)\n",
    "\n",
    "    # Now count the percentages \n",
    "    \n",
    "    error_per_1.append(high_error1/sample_cnt)\n",
    "    error_per_2.append(high_error2/sample_cnt)\n",
    "    error_per_3.append(high_error3/sample_cnt)\n",
    "    error_per_baseline.append(high_error_baseline/ sample_cnt)\n",
    "\n",
    "\n",
    "    # Useful for plotting later on\n",
    "    threshold_list.append(threshold)\n",
    "    sample_cnt_list.append(sample_cnt)\n",
    "    #print(\"The sample count is \" + str(sample_cnt))\n",
    "\n"
   ],
   "outputs": [],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "source": [
    "#Note that this is for 30 degrees since we start from 10.\n",
    "print(sample_cnt_list[20])\n",
    "\n",
    "\n",
    "print(losses_mean + losses_std)\n",
    "print(losses_mean + 2*losses_std)\n",
    "\n"
   ],
   "outputs": [
    {
     "output_type": "stream",
     "name": "stdout",
     "text": [
      "8853\n",
      "5.70443471405585\n",
      "8.206063898339105\n"
     ]
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "source": [
    "#Plotting the number of images against the number of values outside a given range \n",
    "\n",
    "plt.close('all')\n",
    "\n",
    "fig, ax = plt.subplots(tight_layout=True)\n",
    "\n",
    "p1,  = ax.plot(threshold_list[0:20], sample_cnt_list[0:20], color='blue', marker='o', label='Number of samples with high error/Total number of samples outside range')\n",
    "\n",
    "ax.set_title(\"Number of images vs Pitch,Yaw Threshold\")\n",
    "ax.set_xlabel(\"Threshold of the Pitch,Yaw range(In degrees)\")\n",
    "ax.set_ylabel(\"Number of iamges ouside threshold\")\n",
    "\n",
    "ax.ticklabel_format(style='plain')\n",
    "\n",
    "ax.grid(True)\n",
    "\n",
    "fig.savefig('/bigpool/fachpraktikum2021/group5/dir_srinivas/STED-gaze/Gaze_Estimation/plots/' + 'N_Images_vs_PitchYaw_threshold.jpg',\n",
    "            )"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ],
      "image/png": "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"
     },
     "metadata": {
      "needs_background": "light"
     }
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "source": [
    "\n",
    "# Plotting the ratio of samples displaying high error vs threshold\n",
    "# This is done so since we want to find the max threshold beyond which we expereince very high error. \n",
    "\n",
    "plt.close('all')\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "p1,  = ax.plot(threshold_list[0:20], error_per_1[0:20], color='blue', marker='o', label='Number of samples with high error/\\nTotal number of samples outside range')\n",
    "#p2,  = plt.plot(threshold_list[0:20], error_per_2[0:20], color='red', marker='o', label='Number of samples with high error/Total number of samples outside range')\n",
    "\n",
    "#p4,  = plt.plot(threshold_list[0:20], error_per_baseline[0:20], color='black', marker='x', label='Number of samples with high error/Total number of samples outside range')\n",
    "\n",
    "#ax.set_yticks(np.arange(0, 1, 0.1))\n",
    "\n",
    "\n",
    "ax.set_title(\"Error Ratio vs Pitch/Yaw Angle Threshold\")\n",
    "ax.set_xlabel(\"Threshold of the Pitch/Yaw range(In degrees)\")\n",
    "ax.set_ylabel(\"Ration of N(high error)/N(outside range)\")\n",
    "\n",
    "#ax.legend(handles=[p1], title='Legend', bbox_to_anchor=(1.05, 1), loc='upper left')\n",
    "\n",
    "ax.grid(True)\n",
    "\n",
    "fig.savefig('/bigpool/fachpraktikum2021/group5/dir_srinivas/STED-gaze/Gaze_Estimation/plots/' + 'Error_Ratio_vs_PitchYaw_threshold.jpg',\n",
    "            )"
   ],
   "outputs": [
    {
     "output_type": "display_data",
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ],
      "image/png": "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"
     },
     "metadata": {
      "needs_background": "light"
     }
    }
   ],
   "metadata": {}
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "source": [],
   "outputs": [],
   "metadata": {}
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "0a7e8266d4960bf55ec4202778d20d6fb4fd9d14c12b2b3e2041a81f30d16324"
  },
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3.8.0 64-bit ('venv38': venv)"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.0"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}