Flavia py files/classify_leaves_flavia.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plant Leaf Classification\n",
    "\n",
    "Applying machine learning models for classification of plant leaf images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Importing necessary libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import string"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reading the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "dataset = pd.read_csv(\"Flavia_features.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>area</th>\n",
       "      <th>perimeter</th>\n",
       "      <th>physiological_length</th>\n",
       "      <th>physiological_width</th>\n",
       "      <th>aspect_ratio</th>\n",
       "      <th>rectangularity</th>\n",
       "      <th>circularity</th>\n",
       "      <th>mean_r</th>\n",
       "      <th>mean_g</th>\n",
       "      <th>mean_b</th>\n",
       "      <th>stddev_r</th>\n",
       "      <th>stddev_g</th>\n",
       "      <th>stddev_b</th>\n",
       "      <th>contrast</th>\n",
       "      <th>correlation</th>\n",
       "      <th>inverse_difference_moments</th>\n",
       "      <th>entropy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>197484.0</td>\n",
       "      <td>3479.036038</td>\n",
       "      <td>1416</td>\n",
       "      <td>759</td>\n",
       "      <td>1.865613</td>\n",
       "      <td>5.442183</td>\n",
       "      <td>61.289480</td>\n",
       "      <td>6.395667</td>\n",
       "      <td>13.643413</td>\n",
       "      <td>4.388007</td>\n",
       "      <td>24.025329</td>\n",
       "      <td>40.200931</td>\n",
       "      <td>21.448412</td>\n",
       "      <td>12.639656</td>\n",
       "      <td>0.997666</td>\n",
       "      <td>0.911738</td>\n",
       "      <td>1.688689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>101248.0</td>\n",
       "      <td>2490.381812</td>\n",
       "      <td>1190</td>\n",
       "      <td>130</td>\n",
       "      <td>9.153846</td>\n",
       "      <td>1.527931</td>\n",
       "      <td>61.255546</td>\n",
       "      <td>7.049316</td>\n",
       "      <td>9.232018</td>\n",
       "      <td>10.876066</td>\n",
       "      <td>33.816205</td>\n",
       "      <td>37.382221</td>\n",
       "      <td>46.479230</td>\n",
       "      <td>8.137424</td>\n",
       "      <td>0.997191</td>\n",
       "      <td>0.944818</td>\n",
       "      <td>1.193795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>86570.5</td>\n",
       "      <td>2290.683327</td>\n",
       "      <td>1095</td>\n",
       "      <td>119</td>\n",
       "      <td>9.201681</td>\n",
       "      <td>1.505189</td>\n",
       "      <td>60.612219</td>\n",
       "      <td>3.434303</td>\n",
       "      <td>6.371511</td>\n",
       "      <td>2.644757</td>\n",
       "      <td>19.975699</td>\n",
       "      <td>29.057374</td>\n",
       "      <td>19.275050</td>\n",
       "      <td>8.553729</td>\n",
       "      <td>0.996610</td>\n",
       "      <td>0.959023</td>\n",
       "      <td>0.848758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>190214.0</td>\n",
       "      <td>2856.479353</td>\n",
       "      <td>1318</td>\n",
       "      <td>254</td>\n",
       "      <td>5.188976</td>\n",
       "      <td>1.759976</td>\n",
       "      <td>42.896287</td>\n",
       "      <td>7.670415</td>\n",
       "      <td>13.303599</td>\n",
       "      <td>6.049157</td>\n",
       "      <td>28.822885</td>\n",
       "      <td>40.221845</td>\n",
       "      <td>26.948604</td>\n",
       "      <td>8.440041</td>\n",
       "      <td>0.998419</td>\n",
       "      <td>0.914331</td>\n",
       "      <td>1.673915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>227727.0</td>\n",
       "      <td>2917.248904</td>\n",
       "      <td>1324</td>\n",
       "      <td>286</td>\n",
       "      <td>4.629371</td>\n",
       "      <td>1.662798</td>\n",
       "      <td>37.370804</td>\n",
       "      <td>8.992028</td>\n",
       "      <td>16.671171</td>\n",
       "      <td>6.294281</td>\n",
       "      <td>30.967158</td>\n",
       "      <td>45.040192</td>\n",
       "      <td>28.595326</td>\n",
       "      <td>8.641438</td>\n",
       "      <td>0.998568</td>\n",
       "      <td>0.898644</td>\n",
       "      <td>1.968082</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0      area    perimeter  physiological_length  \\\n",
       "0           0  197484.0  3479.036038                  1416   \n",
       "1           0  101248.0  2490.381812                  1190   \n",
       "2           0   86570.5  2290.683327                  1095   \n",
       "3           0  190214.0  2856.479353                  1318   \n",
       "4           0  227727.0  2917.248904                  1324   \n",
       "\n",
       "   physiological_width  aspect_ratio  rectangularity  circularity    mean_r  \\\n",
       "0                  759      1.865613        5.442183    61.289480  6.395667   \n",
       "1                  130      9.153846        1.527931    61.255546  7.049316   \n",
       "2                  119      9.201681        1.505189    60.612219  3.434303   \n",
       "3                  254      5.188976        1.759976    42.896287  7.670415   \n",
       "4                  286      4.629371        1.662798    37.370804  8.992028   \n",
       "\n",
       "      mean_g     mean_b   stddev_r   stddev_g   stddev_b   contrast  \\\n",
       "0  13.643413   4.388007  24.025329  40.200931  21.448412  12.639656   \n",
       "1   9.232018  10.876066  33.816205  37.382221  46.479230   8.137424   \n",
       "2   6.371511   2.644757  19.975699  29.057374  19.275050   8.553729   \n",
       "3  13.303599   6.049157  28.822885  40.221845  26.948604   8.440041   \n",
       "4  16.671171   6.294281  30.967158  45.040192  28.595326   8.641438   \n",
       "\n",
       "   correlation  inverse_difference_moments   entropy  \n",
       "0     0.997666                    0.911738  1.688689  \n",
       "1     0.997191                    0.944818  1.193795  \n",
       "2     0.996610                    0.959023  0.848758  \n",
       "3     0.998419                    0.914331  1.673915  \n",
       "4     0.998568                    0.898644  1.968082  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "pandas.core.frame.DataFrame"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "maindir = r'D:\\Projects\\Plant Leaf Identification'\n",
    "ds_path = maindir + \"\\\\Flavia leaves dataset\"\n",
    "img_files = os.listdir(ds_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating target labels\n",
    "\n",
    "Breakpoints are used alongside the image file to create a vector of target labels. The breakpoints are specified in Flavia leaves dataset website."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "breakpoints = [1001,1059,1060,1122,1552,1616,1123,1194,1195,1267,1268,1323,1324,1385,1386,1437,1497,1551,1438,1496,2001,2050,2051,2113,2114,2165,2166,2230,2231,2290,2291,2346,2347,2423,2424,2485,2486,2546,2547,2612,2616,2675,3001,3055,3056,3110,3111,3175,3176,3229,3230,3281,3282,3334,3335,3389,3390,3446,3447,3510,3511,3563,3566,3621]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "target_list = []\n",
    "for file in img_files:\n",
    "    target_num = int(file.split(\".\")[0])\n",
    "    flag = 0\n",
    "    i = 0 \n",
    "    for i in range(0,len(breakpoints),2):\n",
    "        if((target_num >= breakpoints[i]) and (target_num <= breakpoints[i+1])):\n",
    "            flag = 1\n",
    "            break\n",
    "    if(flag==1):\n",
    "        target = int((i/2))\n",
    "        target_list.append(target)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0,  0,  0, ..., 31, 31, 31])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = np.array(target_list)\n",
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = dataset.iloc[:,1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>area</th>\n",
       "      <th>perimeter</th>\n",
       "      <th>physiological_length</th>\n",
       "      <th>physiological_width</th>\n",
       "      <th>aspect_ratio</th>\n",
       "      <th>rectangularity</th>\n",
       "      <th>circularity</th>\n",
       "      <th>mean_r</th>\n",
       "      <th>mean_g</th>\n",
       "      <th>mean_b</th>\n",
       "      <th>stddev_r</th>\n",
       "      <th>stddev_g</th>\n",
       "      <th>stddev_b</th>\n",
       "      <th>contrast</th>\n",
       "      <th>correlation</th>\n",
       "      <th>inverse_difference_moments</th>\n",
       "      <th>entropy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>197484.0</td>\n",
       "      <td>3479.036038</td>\n",
       "      <td>1416</td>\n",
       "      <td>759</td>\n",
       "      <td>1.865613</td>\n",
       "      <td>5.442183</td>\n",
       "      <td>61.289480</td>\n",
       "      <td>6.395667</td>\n",
       "      <td>13.643413</td>\n",
       "      <td>4.388007</td>\n",
       "      <td>24.025329</td>\n",
       "      <td>40.200931</td>\n",
       "      <td>21.448412</td>\n",
       "      <td>12.639656</td>\n",
       "      <td>0.997666</td>\n",
       "      <td>0.911738</td>\n",
       "      <td>1.688689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>101248.0</td>\n",
       "      <td>2490.381812</td>\n",
       "      <td>1190</td>\n",
       "      <td>130</td>\n",
       "      <td>9.153846</td>\n",
       "      <td>1.527931</td>\n",
       "      <td>61.255546</td>\n",
       "      <td>7.049316</td>\n",
       "      <td>9.232018</td>\n",
       "      <td>10.876066</td>\n",
       "      <td>33.816205</td>\n",
       "      <td>37.382221</td>\n",
       "      <td>46.479230</td>\n",
       "      <td>8.137424</td>\n",
       "      <td>0.997191</td>\n",
       "      <td>0.944818</td>\n",
       "      <td>1.193795</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>86570.5</td>\n",
       "      <td>2290.683327</td>\n",
       "      <td>1095</td>\n",
       "      <td>119</td>\n",
       "      <td>9.201681</td>\n",
       "      <td>1.505189</td>\n",
       "      <td>60.612219</td>\n",
       "      <td>3.434303</td>\n",
       "      <td>6.371511</td>\n",
       "      <td>2.644757</td>\n",
       "      <td>19.975699</td>\n",
       "      <td>29.057374</td>\n",
       "      <td>19.275050</td>\n",
       "      <td>8.553729</td>\n",
       "      <td>0.996610</td>\n",
       "      <td>0.959023</td>\n",
       "      <td>0.848758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>190214.0</td>\n",
       "      <td>2856.479353</td>\n",
       "      <td>1318</td>\n",
       "      <td>254</td>\n",
       "      <td>5.188976</td>\n",
       "      <td>1.759976</td>\n",
       "      <td>42.896287</td>\n",
       "      <td>7.670415</td>\n",
       "      <td>13.303599</td>\n",
       "      <td>6.049157</td>\n",
       "      <td>28.822885</td>\n",
       "      <td>40.221845</td>\n",
       "      <td>26.948604</td>\n",
       "      <td>8.440041</td>\n",
       "      <td>0.998419</td>\n",
       "      <td>0.914331</td>\n",
       "      <td>1.673915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>227727.0</td>\n",
       "      <td>2917.248904</td>\n",
       "      <td>1324</td>\n",
       "      <td>286</td>\n",
       "      <td>4.629371</td>\n",
       "      <td>1.662798</td>\n",
       "      <td>37.370804</td>\n",
       "      <td>8.992028</td>\n",
       "      <td>16.671171</td>\n",
       "      <td>6.294281</td>\n",
       "      <td>30.967158</td>\n",
       "      <td>45.040192</td>\n",
       "      <td>28.595326</td>\n",
       "      <td>8.641438</td>\n",
       "      <td>0.998568</td>\n",
       "      <td>0.898644</td>\n",
       "      <td>1.968082</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       area    perimeter  physiological_length  physiological_width  \\\n",
       "0  197484.0  3479.036038                  1416                  759   \n",
       "1  101248.0  2490.381812                  1190                  130   \n",
       "2   86570.5  2290.683327                  1095                  119   \n",
       "3  190214.0  2856.479353                  1318                  254   \n",
       "4  227727.0  2917.248904                  1324                  286   \n",
       "\n",
       "   aspect_ratio  rectangularity  circularity    mean_r     mean_g     mean_b  \\\n",
       "0      1.865613        5.442183    61.289480  6.395667  13.643413   4.388007   \n",
       "1      9.153846        1.527931    61.255546  7.049316   9.232018  10.876066   \n",
       "2      9.201681        1.505189    60.612219  3.434303   6.371511   2.644757   \n",
       "3      5.188976        1.759976    42.896287  7.670415  13.303599   6.049157   \n",
       "4      4.629371        1.662798    37.370804  8.992028  16.671171   6.294281   \n",
       "\n",
       "    stddev_r   stddev_g   stddev_b   contrast  correlation  \\\n",
       "0  24.025329  40.200931  21.448412  12.639656     0.997666   \n",
       "1  33.816205  37.382221  46.479230   8.137424     0.997191   \n",
       "2  19.975699  29.057374  19.275050   8.553729     0.996610   \n",
       "3  28.822885  40.221845  26.948604   8.440041     0.998419   \n",
       "4  30.967158  45.040192  28.595326   8.641438     0.998568   \n",
       "\n",
       "   inverse_difference_moments   entropy  \n",
       "0                    0.911738  1.688689  \n",
       "1                    0.944818  1.193795  \n",
       "2                    0.959023  0.848758  \n",
       "3                    0.914331  1.673915  \n",
       "4                    0.898644  1.968082  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, 0, 0])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[0:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state = 142)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>area</th>\n",
       "      <th>perimeter</th>\n",
       "      <th>physiological_length</th>\n",
       "      <th>physiological_width</th>\n",
       "      <th>aspect_ratio</th>\n",
       "      <th>rectangularity</th>\n",
       "      <th>circularity</th>\n",
       "      <th>mean_r</th>\n",
       "      <th>mean_g</th>\n",
       "      <th>mean_b</th>\n",
       "      <th>stddev_r</th>\n",
       "      <th>stddev_g</th>\n",
       "      <th>stddev_b</th>\n",
       "      <th>contrast</th>\n",
       "      <th>correlation</th>\n",
       "      <th>inverse_difference_moments</th>\n",
       "      <th>entropy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>816496.5</td>\n",
       "      <td>3599.018794</td>\n",
       "      <td>1101</td>\n",
       "      <td>1083</td>\n",
       "      <td>1.016620</td>\n",
       "      <td>1.460365</td>\n",
       "      <td>15.864044</td>\n",
       "      <td>36.740296</td>\n",
       "      <td>70.312576</td>\n",
       "      <td>25.082371</td>\n",
       "      <td>45.407495</td>\n",
       "      <td>81.167757</td>\n",
       "      <td>34.334657</td>\n",
       "      <td>22.519541</td>\n",
       "      <td>0.997333</td>\n",
       "      <td>0.649635</td>\n",
       "      <td>5.610553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>837</th>\n",
       "      <td>921134.0</td>\n",
       "      <td>4097.732429</td>\n",
       "      <td>1377</td>\n",
       "      <td>1038</td>\n",
       "      <td>1.326590</td>\n",
       "      <td>1.551703</td>\n",
       "      <td>18.229064</td>\n",
       "      <td>34.696545</td>\n",
       "      <td>64.961098</td>\n",
       "      <td>23.699098</td>\n",
       "      <td>46.639467</td>\n",
       "      <td>69.345675</td>\n",
       "      <td>35.581559</td>\n",
       "      <td>92.360837</td>\n",
       "      <td>0.992630</td>\n",
       "      <td>0.560696</td>\n",
       "      <td>7.248010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>358</th>\n",
       "      <td>552379.0</td>\n",
       "      <td>3850.814822</td>\n",
       "      <td>1398</td>\n",
       "      <td>1049</td>\n",
       "      <td>1.332698</td>\n",
       "      <td>2.654884</td>\n",
       "      <td>26.845291</td>\n",
       "      <td>29.863131</td>\n",
       "      <td>49.430173</td>\n",
       "      <td>23.717367</td>\n",
       "      <td>48.482326</td>\n",
       "      <td>77.065204</td>\n",
       "      <td>39.447077</td>\n",
       "      <td>23.181824</td>\n",
       "      <td>0.996204</td>\n",
       "      <td>0.751559</td>\n",
       "      <td>4.221778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>962</th>\n",
       "      <td>38454.5</td>\n",
       "      <td>3035.410698</td>\n",
       "      <td>1176</td>\n",
       "      <td>822</td>\n",
       "      <td>1.430657</td>\n",
       "      <td>25.138072</td>\n",
       "      <td>239.600518</td>\n",
       "      <td>4.282973</td>\n",
       "      <td>5.213604</td>\n",
       "      <td>4.262034</td>\n",
       "      <td>27.772276</td>\n",
       "      <td>30.832106</td>\n",
       "      <td>28.934080</td>\n",
       "      <td>6.313081</td>\n",
       "      <td>0.993080</td>\n",
       "      <td>0.971889</td>\n",
       "      <td>0.634766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1101</th>\n",
       "      <td>941172.0</td>\n",
       "      <td>3719.580197</td>\n",
       "      <td>1303</td>\n",
       "      <td>987</td>\n",
       "      <td>1.320162</td>\n",
       "      <td>1.366446</td>\n",
       "      <td>14.700051</td>\n",
       "      <td>43.594426</td>\n",
       "      <td>72.467561</td>\n",
       "      <td>32.378145</td>\n",
       "      <td>47.408525</td>\n",
       "      <td>73.463707</td>\n",
       "      <td>37.026025</td>\n",
       "      <td>11.663863</td>\n",
       "      <td>0.998818</td>\n",
       "      <td>0.635070</td>\n",
       "      <td>6.042620</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          area    perimeter  physiological_length  physiological_width  \\\n",
       "147   816496.5  3599.018794                  1101                 1083   \n",
       "837   921134.0  4097.732429                  1377                 1038   \n",
       "358   552379.0  3850.814822                  1398                 1049   \n",
       "962    38454.5  3035.410698                  1176                  822   \n",
       "1101  941172.0  3719.580197                  1303                  987   \n",
       "\n",
       "      aspect_ratio  rectangularity  circularity     mean_r     mean_g  \\\n",
       "147       1.016620        1.460365    15.864044  36.740296  70.312576   \n",
       "837       1.326590        1.551703    18.229064  34.696545  64.961098   \n",
       "358       1.332698        2.654884    26.845291  29.863131  49.430173   \n",
       "962       1.430657       25.138072   239.600518   4.282973   5.213604   \n",
       "1101      1.320162        1.366446    14.700051  43.594426  72.467561   \n",
       "\n",
       "         mean_b   stddev_r   stddev_g   stddev_b   contrast  correlation  \\\n",
       "147   25.082371  45.407495  81.167757  34.334657  22.519541     0.997333   \n",
       "837   23.699098  46.639467  69.345675  35.581559  92.360837     0.992630   \n",
       "358   23.717367  48.482326  77.065204  39.447077  23.181824     0.996204   \n",
       "962    4.262034  27.772276  30.832106  28.934080   6.313081     0.993080   \n",
       "1101  32.378145  47.408525  73.463707  37.026025  11.663863     0.998818   \n",
       "\n",
       "      inverse_difference_moments   entropy  \n",
       "147                     0.649635  5.610553  \n",
       "837                     0.560696  7.248010  \n",
       "358                     0.751559  4.221778  \n",
       "962                     0.971889  0.634766  \n",
       "1101                    0.635070  6.042620  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 3, 13,  6, 16, 18])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[0:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature Scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sc_X = StandardScaler()\n",
    "X_train = sc_X.fit_transform(X_train)\n",
    "X_test = sc_X.transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.83193192, -0.17692354, -0.90777284,  0.86491733, -0.9528451 ,\n",
       "        -0.33789935, -0.42611352,  0.01198242,  0.22801441, -0.06515913,\n",
       "        -0.03038517,  0.66513894, -0.17255204, -0.17456558,  0.27446021,\n",
       "        -0.97141035,  0.72845056],\n",
       "       [ 1.22949107,  0.46823763,  0.25932835,  0.67383324, -0.45412004,\n",
       "        -0.31773827, -0.37146523, -0.04156278,  0.10823179, -0.10299271,\n",
       "         0.04545738, -0.01454148, -0.11043539,  2.3536019 , -1.92405659,\n",
       "        -1.86068666,  1.74156049]])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0:2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 3, 13])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[0:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Applying SVM classifier model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn import svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = svm.SVC()\n",
    "clf.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.79755671902268765"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.accuracy_score(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          0       0.83      0.28      0.42        18\n",
      "          1       0.65      0.72      0.68        18\n",
      "          2       0.90      0.86      0.88        22\n",
      "          3       0.81      0.96      0.88        27\n",
      "          4       0.96      1.00      0.98        24\n",
      "          5       0.89      1.00      0.94        16\n",
      "          6       0.73      0.95      0.83        20\n",
      "          7       0.83      0.71      0.77        14\n",
      "          8       0.35      0.50      0.41        12\n",
      "          9       0.93      0.93      0.93        15\n",
      "         10       0.65      0.94      0.77        18\n",
      "         11       0.92      0.61      0.73        18\n",
      "         12       0.58      0.78      0.67        18\n",
      "         13       0.94      0.71      0.81        24\n",
      "         14       0.60      0.46      0.52        13\n",
      "         15       0.44      0.22      0.30        18\n",
      "         16       1.00      1.00      1.00        26\n",
      "         17       0.94      1.00      0.97        17\n",
      "         18       0.94      0.81      0.87        21\n",
      "         19       0.82      0.93      0.87        15\n",
      "         20       0.62      0.93      0.74        14\n",
      "         21       0.62      0.87      0.72        15\n",
      "         22       1.00      0.93      0.96        14\n",
      "         23       0.70      0.88      0.78        16\n",
      "         24       0.92      0.69      0.79        16\n",
      "         25       0.75      0.69      0.72        13\n",
      "         26       0.93      0.93      0.93        15\n",
      "         27       0.86      0.86      0.86        22\n",
      "         28       0.62      0.71      0.67        14\n",
      "         29       0.85      0.81      0.83        21\n",
      "         30       0.91      0.67      0.77        15\n",
      "         31       0.95      0.79      0.86        24\n",
      "\n",
      "avg / total       0.81      0.80      0.79       573\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Performing parameter tuning of the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "parameters = [{'kernel': ['rbf'],\n",
    "               'gamma': [1e-4, 1e-3, 0.01, 0.1, 0.2, 0.5],\n",
    "               'C': [1, 10, 100, 1000]},\n",
    "              {'kernel': ['linear'], 'C': [1, 10, 100, 1000]}\n",
    "             ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False),\n",
       "       fit_params={}, iid=True, n_jobs=1,\n",
       "       param_grid=[{'kernel': ['rbf'], 'gamma': [0.0001, 0.001, 0.01, 0.1, 0.2, 0.5], 'C': [1, 10, 100, 1000]}, {'kernel': ['linear'], 'C': [1, 10, 100, 1000]}],\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score=True,\n",
       "       scoring=None, verbose=0)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_clf = GridSearchCV(svm.SVC(decision_function_shape='ovr'), parameters, cv=5)\n",
    "svm_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'C': 100, 'kernel': 'linear'}"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_clf.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.076 (+/-0.005) for {'C': 1, 'gamma': 0.0001, 'kernel': 'rbf'}\n",
      "0.198 (+/-0.023) for {'C': 1, 'gamma': 0.001, 'kernel': 'rbf'}\n",
      "0.603 (+/-0.061) for {'C': 1, 'gamma': 0.01, 'kernel': 'rbf'}\n",
      "0.833 (+/-0.048) for {'C': 1, 'gamma': 0.1, 'kernel': 'rbf'}\n",
      "0.843 (+/-0.039) for {'C': 1, 'gamma': 0.2, 'kernel': 'rbf'}\n",
      "0.859 (+/-0.035) for {'C': 1, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.199 (+/-0.021) for {'C': 10, 'gamma': 0.0001, 'kernel': 'rbf'}\n",
      "0.602 (+/-0.067) for {'C': 10, 'gamma': 0.001, 'kernel': 'rbf'}\n",
      "0.839 (+/-0.035) for {'C': 10, 'gamma': 0.01, 'kernel': 'rbf'}\n",
      "0.888 (+/-0.027) for {'C': 10, 'gamma': 0.1, 'kernel': 'rbf'}\n",
      "0.882 (+/-0.014) for {'C': 10, 'gamma': 0.2, 'kernel': 'rbf'}\n",
      "0.872 (+/-0.017) for {'C': 10, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.599 (+/-0.068) for {'C': 100, 'gamma': 0.0001, 'kernel': 'rbf'}\n",
      "0.841 (+/-0.043) for {'C': 100, 'gamma': 0.001, 'kernel': 'rbf'}\n",
      "0.891 (+/-0.027) for {'C': 100, 'gamma': 0.01, 'kernel': 'rbf'}\n",
      "0.888 (+/-0.008) for {'C': 100, 'gamma': 0.1, 'kernel': 'rbf'}\n",
      "0.875 (+/-0.012) for {'C': 100, 'gamma': 0.2, 'kernel': 'rbf'}\n",
      "0.874 (+/-0.018) for {'C': 100, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.843 (+/-0.043) for {'C': 1000, 'gamma': 0.0001, 'kernel': 'rbf'}\n",
      "0.891 (+/-0.024) for {'C': 1000, 'gamma': 0.001, 'kernel': 'rbf'}\n",
      "0.903 (+/-0.013) for {'C': 1000, 'gamma': 0.01, 'kernel': 'rbf'}\n",
      "0.888 (+/-0.022) for {'C': 1000, 'gamma': 0.1, 'kernel': 'rbf'}\n",
      "0.877 (+/-0.015) for {'C': 1000, 'gamma': 0.2, 'kernel': 'rbf'}\n",
      "0.875 (+/-0.020) for {'C': 1000, 'gamma': 0.5, 'kernel': 'rbf'}\n",
      "0.873 (+/-0.036) for {'C': 1, 'kernel': 'linear'}\n",
      "0.896 (+/-0.015) for {'C': 10, 'kernel': 'linear'}\n",
      "0.904 (+/-0.016) for {'C': 100, 'kernel': 'linear'}\n",
      "0.903 (+/-0.027) for {'C': 1000, 'kernel': 'linear'}\n"
     ]
    }
   ],
   "source": [
    "means = svm_clf.cv_results_['mean_test_score']\n",
    "stds = svm_clf.cv_results_['std_test_score']\n",
    "for mean, std, params in zip(means, stds, svm_clf.cv_results_['params']):\n",
    "    print(\"%0.3f (+/-%0.03f) for %r\" % (mean, std * 2, params))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred_svm = svm_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.90052356020942403"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.accuracy_score(y_test, y_pred_svm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          0       0.93      0.78      0.85        18\n",
      "          1       0.82      0.78      0.80        18\n",
      "          2       1.00      0.91      0.95        22\n",
      "          3       0.93      0.96      0.95        27\n",
      "          4       1.00      1.00      1.00        24\n",
      "          5       0.94      1.00      0.97        16\n",
      "          6       0.73      0.95      0.83        20\n",
      "          7       0.87      0.93      0.90        14\n",
      "          8       0.54      0.58      0.56        12\n",
      "          9       1.00      1.00      1.00        15\n",
      "         10       0.94      0.89      0.91        18\n",
      "         11       0.72      0.72      0.72        18\n",
      "         12       0.94      0.89      0.91        18\n",
      "         13       0.81      0.88      0.84        24\n",
      "         14       0.92      0.85      0.88        13\n",
      "         15       0.81      0.72      0.76        18\n",
      "         16       1.00      1.00      1.00        26\n",
      "         17       0.94      1.00      0.97        17\n",
      "         18       0.88      1.00      0.93        21\n",
      "         19       0.88      0.93      0.90        15\n",
      "         20       1.00      1.00      1.00        14\n",
      "         21       0.88      1.00      0.94        15\n",
      "         22       1.00      1.00      1.00        14\n",
      "         23       0.84      1.00      0.91        16\n",
      "         24       0.93      0.81      0.87        16\n",
      "         25       0.90      0.69      0.78        13\n",
      "         26       0.94      1.00      0.97        15\n",
      "         27       1.00      0.86      0.93        22\n",
      "         28       0.93      0.93      0.93        14\n",
      "         29       0.90      0.90      0.90        21\n",
      "         30       0.92      0.73      0.81        15\n",
      "         31       0.96      0.92      0.94        24\n",
      "\n",
      "avg / total       0.90      0.90      0.90       573\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(metrics.classification_report(y_test, y_pred_svm))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dimensionality Reduction using PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pca = PCA()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PCA(copy=True, iterated_power='auto', n_components=None, random_state=None,\n",
       "  svd_solver='auto', tol=0.0, whiten=False)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca.fit(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  9.99992897e-01,   6.41437085e-06,   3.92824599e-07,\n",
       "         2.08042088e-07,   6.15522995e-08,   1.38782111e-08,\n",
       "         8.94933506e-09,   2.51990655e-09,   8.66436338e-10,\n",
       "         2.10139778e-10,   1.07067132e-10,   1.32410441e-11,\n",
       "         2.22566990e-12,   1.41983601e-12,   5.16456146e-13,\n",
       "         3.39145452e-15,   1.00669705e-17])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "var= pca.explained_variance_ratio_\n",
    "var"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1821e605d68>]"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD8CAYAAAB5Pm/hAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSYSsgmtNKgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1821e556a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "var1=np.cumsum(np.round(pca.explained_variance_ratio_, decimals=4)*100)\n",
    "plt.plot(var1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Testing with mobile captured leaves which are not classified"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import cv2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def bg_sub(filename):\n",
    "    test_img_path = '..\\\\mobile captures\\\\' + filename\n",
    "    main_img = cv2.imread(test_img_path)\n",
    "    img = cv2.cvtColor(main_img, cv2.COLOR_BGR2RGB)\n",
    "    resized_image = cv2.resize(img, (1600, 1200))\n",
    "    size_y,size_x,_ = img.shape\n",
    "    gs = cv2.cvtColor(resized_image,cv2.COLOR_RGB2GRAY)\n",
    "    blur = cv2.GaussianBlur(gs, (55,55),0)\n",
    "    ret_otsu,im_bw_otsu = cv2.threshold(blur,0,255,cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)\n",
    "    kernel = np.ones((50,50),np.uint8)\n",
    "    closing = cv2.morphologyEx(im_bw_otsu, cv2.MORPH_CLOSE, kernel)\n",
    "    \n",
    "    ret_image, contours, hierarchy = cv2.findContours(closing,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)\n",
    "    \n",
    "    contains = []\n",
    "    y_ri,x_ri, _ = resized_image.shape\n",
    "    for cc in contours:\n",
    "        yn = cv2.pointPolygonTest(cc,(x_ri//2,y_ri//2),False)\n",
    "        contains.append(yn)\n",
    "\n",
    "    val = [contains.index(temp) for temp in contains if temp>0]\n",
    "    index = val[0]\n",
    "    \n",
    "    black_img = np.empty([1200,1600,3],dtype=np.uint8)\n",
    "    black_img.fill(0)\n",
    "    \n",
    "    cnt = contours[index]\n",
    "    mask = cv2.drawContours(black_img, [cnt] , 0, (255,255,255), -1)\n",
    "    \n",
    "    maskedImg = cv2.bitwise_and(resized_image, mask)\n",
    "    white_pix = [255,255,255]\n",
    "    black_pix = [0,0,0]\n",
    "    \n",
    "    final_img = maskedImg\n",
    "    h,w,channels = final_img.shape\n",
    "    for x in range(0,w):\n",
    "        for y in range(0,h):\n",
    "            channels_xy = final_img[y,x]\n",
    "            if all(channels_xy == black_pix):\n",
    "                final_img[y,x] = white_pix\n",
    "    \n",
    "    return final_img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "filename = 'Test.jpg' \n",
    "bg_rem_img = bg_sub(filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x182202eaf60>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAAD8CAYAAADUv3dIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x182202c3198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(bg_rem_img)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import mahotas as mt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def feature_extract(img):\n",
    "    names = ['area','perimeter','pysiological_length','pysiological_width','aspect_ratio','rectangularity','circularity', \\\n",
    "             'mean_r','mean_g','mean_b','stddev_r','stddev_g','stddev_b', \\\n",
    "             'contrast','correlation','inverse_difference_moments','entropy'\n",
    "            ]\n",
    "    df = pd.DataFrame([], columns=names)\n",
    "\n",
    "    #Preprocessing\n",
    "    gs = cv2.cvtColor(img,cv2.COLOR_RGB2GRAY)\n",
    "    blur = cv2.GaussianBlur(gs, (25,25),0)\n",
    "    ret_otsu,im_bw_otsu = cv2.threshold(blur,0,255,cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)\n",
    "    kernel = np.ones((50,50),np.uint8)\n",
    "    closing = cv2.morphologyEx(im_bw_otsu, cv2.MORPH_CLOSE, kernel)\n",
    "\n",
    "    #Shape features\n",
    "    image, contours, _ = cv2.findContours(closing,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)\n",
    "    cnt = contours[0]\n",
    "    M = cv2.moments(cnt)\n",
    "    area = cv2.contourArea(cnt)\n",
    "    perimeter = cv2.arcLength(cnt,True)\n",
    "    x,y,w,h = cv2.boundingRect(cnt)\n",
    "    aspect_ratio = float(w)/h\n",
    "    rectangularity = w*h/area\n",
    "    circularity = ((perimeter)**2)/area\n",
    "\n",
    "    #Color features\n",
    "    red_channel = img[:,:,0]\n",
    "    green_channel = img[:,:,1]\n",
    "    blue_channel = img[:,:,2]\n",
    "    blue_channel[blue_channel == 255] = 0\n",
    "    green_channel[green_channel == 255] = 0\n",
    "    red_channel[red_channel == 255] = 0\n",
    "\n",
    "    red_mean = np.mean(red_channel)\n",
    "    green_mean = np.mean(green_channel)\n",
    "    blue_mean = np.mean(blue_channel)\n",
    "\n",
    "    red_std = np.std(red_channel)\n",
    "    green_std = np.std(green_channel)\n",
    "    blue_std = np.std(blue_channel)\n",
    "\n",
    "    #Texture features\n",
    "    textures = mt.features.haralick(gs)\n",
    "    ht_mean = textures.mean(axis=0)\n",
    "    contrast = ht_mean[1]\n",
    "    correlation = ht_mean[2]\n",
    "    inverse_diff_moments = ht_mean[4]\n",
    "    entropy = ht_mean[8]\n",
    "\n",
    "    vector = [area,perimeter,w,h,aspect_ratio,rectangularity,circularity,\\\n",
    "              red_mean,green_mean,blue_mean,red_std,green_std,blue_std,\\\n",
    "              contrast,correlation,inverse_diff_moments,entropy\n",
    "             ]\n",
    "\n",
    "    df_temp = pd.DataFrame([vector],columns=names)\n",
    "    df = df.append(df_temp)\n",
    "    \n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>area</th>\n",
       "      <th>perimeter</th>\n",
       "      <th>pysiological_length</th>\n",
       "      <th>pysiological_width</th>\n",
       "      <th>aspect_ratio</th>\n",
       "      <th>rectangularity</th>\n",
       "      <th>circularity</th>\n",
       "      <th>mean_r</th>\n",
       "      <th>mean_g</th>\n",
       "      <th>mean_b</th>\n",
       "      <th>stddev_r</th>\n",
       "      <th>stddev_g</th>\n",
       "      <th>stddev_b</th>\n",
       "      <th>contrast</th>\n",
       "      <th>correlation</th>\n",
       "      <th>inverse_difference_moments</th>\n",
       "      <th>entropy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>153912.5</td>\n",
       "      <td>1581.907845</td>\n",
       "      <td>507</td>\n",
       "      <td>486</td>\n",
       "      <td>1.04321</td>\n",
       "      <td>1.600923</td>\n",
       "      <td>16.258799</td>\n",
       "      <td>3.72212</td>\n",
       "      <td>5.285164</td>\n",
       "      <td>1.348841</td>\n",
       "      <td>13.059666</td>\n",
       "      <td>18.114367</td>\n",
       "      <td>5.464123</td>\n",
       "      <td>14.528435</td>\n",
       "      <td>0.99761</td>\n",
       "      <td>0.951879</td>\n",
       "      <td>1.088378</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       area    perimeter pysiological_length pysiological_width  aspect_ratio  \\\n",
       "0  153912.5  1581.907845                 507                486       1.04321   \n",
       "\n",
       "   rectangularity  circularity   mean_r    mean_g    mean_b   stddev_r  \\\n",
       "0        1.600923    16.258799  3.72212  5.285164  1.348841  13.059666   \n",
       "\n",
       "    stddev_g  stddev_b   contrast  correlation  inverse_difference_moments  \\\n",
       "0  18.114367  5.464123  14.528435      0.99761                    0.951879   \n",
       "\n",
       "    entropy  \n",
       "0  1.088378  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features_of_img = feature_extract(bg_rem_img)\n",
    "features_of_img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaled_features = sc_X.transform(features_of_img)\n",
    "print(scaled_features)\n",
    "# y_pred_mobile = svm_clf.predict(features_of_img)\n",
    "y_pred_mobile = svm_clf.predict(scaled_features)\n",
    "y_pred_mobile[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Crape myrtle, Crepe myrtle'"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "common_names = ['pubescent bamboo','Chinese horse chestnut','Anhui Barberry', \\\n",
    "                'Chinese redbud','true indigo','Japanese maple','Nanmu',' castor aralia', \\\n",
    "                'Chinese cinnamon','goldenrain tree','Big-fruited Holly','Japanese cheesewood', \\\n",
    "                'wintersweet','camphortree','Japan Arrowwood','sweet osmanthus','deodar','ginkgo, maidenhair tree', \\\n",
    "                'Crape myrtle, Crepe myrtle','oleander','yew plum pine','Japanese Flowering Cherry','Glossy Privet',\\\n",
    "                'Chinese Toon','peach','Ford Woodlotus','trident maple','Beales barberry','southern magnolia',\\\n",
    "                'Canadian poplar','Chinese tulip tree','tangerine'\n",
    "               ]\n",
    "common_names[y_pred_mobile[0]]"
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}