Metrics-Calculator.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "TEST_SCREENDIM = [1920, 1200]\n",
    "\n",
    "cellx = 30 \n",
    "celly = 40 \n",
    "\n",
    "#dummary fixation coordinates\n",
    "x = np.random.randint(1,1900,(100,1))\n",
    "y = np.random.randint(1,1200,(100,1))\n",
    "duration = np.random.randint(50,3000,(100,1))\n",
    "\n",
    "\n",
    "dt =  np.zeros((100,3))\n",
    "dt[:,:1] = x\n",
    "dt[:,1:2] =  y \n",
    "dt[:,2:3] = duration\n",
    "\n",
    "#Dummy Gaze values\n",
    "GAZE_ARRAY = np.array([\n",
    "    [41, 41, 0],\n",
    "    [80, 80, 0],\n",
    "    [501, 501, 0],\n",
    "    [100, 200, 0],\n",
    "    [1002, 1002, 0],\n",
    "    [400, 400, 0],\n",
    "    [200, 200, 0],\n",
    "    [250, 250, 0],\n",
    "    [540, 540, 0]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ConvexHullArea:  5434.45796823519\n",
      "SpatialDensity:  0.050520833333333334\n",
      "NNI Metric:  0.9214022183914532\n"
     ]
    }
   ],
   "source": [
    "from eyesmetriccalculator import EyesMetricCalculator\n",
    "\n",
    "ec = EyesMetricCalculator(dt,GAZE_ARRAY,TEST_SCREENDIM)\n",
    "\n",
    "print('ConvexHullArea: ' ,ec.convexHull('area').compute() )\n",
    "print('SpatialDensity: ' , ec.spatialDensity(cellx,celly).compute())\n",
    "print('NNI Metric: ' , ec.NNI().compute())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# AOI and Transition Matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The Gaze Transition Matrix\n",
      "              aoi_poly1  aoi_circle2  aoi_poly2\n",
      "aoi_poly1          2.0          2.0        0.0\n",
      "aoi_circle2        1.0          0.0        0.0\n",
      "aoi_poly2          0.0          0.0        0.0 \n",
      "\n",
      "Number of Points outside all AOIs : 2\n",
      "Probability of having Points in aoi_poly1:  0.5555555555555556\n",
      "Probability of having Points in aoi_circle2:  0.2222222222222222\n",
      "Probability of having Points in aoi_poly2:  0.0\n",
      "Plot representation\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from transition_matrix import *\n",
    "\n",
    "#Setting important variables\n",
    "TEST_RADIUS = 100\n",
    "TEST_VERTICES_1 = [[ 20,20] , [400,250], [40,400]]\n",
    "TEST_VERTICES_2 = [[800,800], [1000,800], [700,650]]\n",
    "\n",
    "\n",
    "#An AOI dict having all the important areas of interest that will be used to calculate different metrics\n",
    "TEST_AOI_DICT = {\n",
    "                'aoi_poly1': PolyAOI(TEST_SCREENDIM,TEST_VERTICES_1),\n",
    "                'aoi_circle2': CircleAOI(TEST_SCREENDIM, TEST_RADIUS, center=[500,500]),\n",
    "                'aoi_poly2': PolyAOI(TEST_SCREENDIM,TEST_VERTICES_2)\n",
    "                }\n",
    "\n",
    "gz = GazeTransitions(TEST_SCREENDIM, TEST_AOI_DICT, GAZE_ARRAY)\n",
    "print('The Gaze Transition Matrix\\n',gz.get_transition_matrix(),'\\n')\n",
    "print('Number of Points outside all AOIs :',gz.points_outside_aoi)\n",
    "for n,a in TEST_AOI_DICT.items():\n",
    "    print('Probability of having Points in '+ n +': ',gz.get_prob_aoi(n))\n",
    "print('Plot representation')\n",
    "gz.plot_all(annotate_points=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Entropy Metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For entropy two metrics will be calculated as mentioned in the Paper Krejtz et al., 2014\n",
    "\n",
    "Gaze Transition Entropy:  \n",
    "\n",
    "\\begin{equation*}\n",
    "    H_t =- \\sum_i \\pi_i \\sum_j p_{ij}log p_{ij}\n",
    "\\end{equation*}\n",
    "\n",
    "Gaze Stationary Entropy:\n",
    "\n",
    "\\begin{equation*}\n",
    "    H_s =- \\sum_i \\pi_i log \\pi_i\n",
    "\\end{equation*}\n",
    "\n",
    "where  $p_{ij} = \\frac{n_{ij}}{n}$  and $\\pi_i = \\frac{n_i}{n} $\n",
    "\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Gaze Transition Entropy:  0.43\n",
      "Gaze Transition Probability Matrix\n",
      "[[0.22222222 0.22222222 0.        ]\n",
      " [0.11111111 0.         0.        ]\n",
      " [0.         0.         0.        ]]\n",
      "Gaze Stationary Entropy:  0.66\n",
      "Gaze Stationary Probability List [0.5555555555555556, 0.2222222222222222, 0.0]\n"
     ]
    }
   ],
   "source": [
    "print('Gaze Transition Entropy: ' , ec.GEntropy(TEST_AOI_DICT,'transition').compute())\n",
    "print('Gaze Transition Probability Matrix')\n",
    "print(ec.GEntropy(TEST_AOI_DICT,'transition').get_tranprobabilities())\n",
    "print('Gaze Stationary Entropy: ' , ec.GEntropy(TEST_AOI_DICT,'stationary').compute())\n",
    "print('Gaze Stationary Probability List',ec.GEntropy(TEST_AOI_DICT,'transition').get_stationary_prob())"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}