Extract_LCD_dataset.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import json\n",
    "import numpy as np\n",
    "from tqdm import tqdm\n",
    "from skimage import io, util, transform, draw, color, img_as_ubyte\n",
    "from preprocess import utils\n",
    "from extract_lcd import extract_display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "DEFAULT_IMAGE_SHAPE = (450, 800, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def segmentation2polygon(segmentation):\n",
    "    if segmentation.ndim == 2:\n",
    "        segmentation = np.array(segmentation[0])\n",
    "    elif segmentation.ndim == 1:\n",
    "        segmentation = np.array(segmentation)\n",
    "    else:\n",
    "        raise ValueError\n",
    "    yy = segmentation[0::2]\n",
    "    xx = segmentation[1::2]\n",
    "    polygon = np.c_[xx,yy]\n",
    "    return polygon\n",
    "\n",
    "    \n",
    "def iou_score(mask_true, mask_pred):\n",
    "    intersection = np.logical_and(mask_true, mask_pred)\n",
    "    union = np.logical_or(mask_true, mask_pred)\n",
    "    iou = np.sum(intersection) / np.sum(union)\n",
    "    return iou"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_iou(image_dir, path_annotation):\n",
    "    \n",
    "    img_names, iou_scores = [], []\n",
    "\n",
    "    with open(path_annotation) as file:\n",
    "        data = json.load(file)\n",
    "        \n",
    "    for img_info, img_annotation in tqdm(zip(data['images'], data['annotations']), desc=\"Progress\"):\n",
    "        # determine image path\n",
    "        img_name = img_info['file_name']\n",
    "        img_path = os.path.join(image_dir, img_name)    \n",
    "        \n",
    "        # read image\n",
    "        img = util.img_as_float64(io.imread(img_path))\n",
    "        _, mask_pred, _ = extract_display(img, shape=DEFAULT_IMAGE_SHAPE, plot=False)\n",
    "        \n",
    "        # retrieve polygon\n",
    "        segmentation = np.array(img_annotation['segmentation'])\n",
    "        polygon = segmentation2polygon(segmentation)\n",
    "        \n",
    "        # compute mask\n",
    "        mask_true = draw.polygon2mask(img.shape[0:2], polygon)\n",
    "        mask_true = transform.resize(mask_true, output_shape=DEFAULT_IMAGE_SHAPE[:2])\n",
    "        mask_true = utils.pad_binary_image(mask_true, dh=20, dw=20)\n",
    "        mask_true = mask_true.astype(np.uint8)\n",
    "        \n",
    "        img_iou = iou_score(mask_true, mask_pred)\n",
    "        # print(f'IoU = {img_iou}')\n",
    "        \n",
    "        img_names.append(img_name)\n",
    "        iou_scores.append(img_iou)\n",
    "    return img_names, iou_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Progress: 50it [12:16, 14.73s/it]\n"
     ]
    }
   ],
   "source": [
    "image_dir_1 = '/home/aboumessouer/Downloads/YUVA_EB_DATASET/RAW_IMAGES/1_Day_Light/'\n",
    "path_annotation_1 = 'annotations/annotations_01_daylight.json'\n",
    "img_names_1, iou_scores_1 = calculate_iou(image_dir_1, path_annotation_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Progress: 48it [10:53, 13.62s/it]\n"
     ]
    }
   ],
   "source": [
    "image_dir_2 = '/home/aboumessouer/Downloads/YUVA_EB_DATASET/RAW_IMAGES/2_Night/'\n",
    "path_annotation_2 = 'annotations/annotations_02_night.json'\n",
    "img_names_2, iou_scores_2 = calculate_iou(image_dir_2, path_annotation_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Progress: 50it [13:28, 16.17s/it]\n"
     ]
    }
   ],
   "source": [
    "image_dir_3 = '/home/aboumessouer/Downloads/YUVA_EB_DATASET/RAW_IMAGES/3_Tilted/'\n",
    "path_annotation_3 = 'annotations/annotations_03_tilted.json'\n",
    "img_names_3, iou_scores_3 = calculate_iou(image_dir_3, path_annotation_3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Progress: 20it [05:21, 16.05s/it]\n"
     ]
    }
   ],
   "source": [
    "image_dir_4 = '/home/aboumessouer/Downloads/YUVA_EB_DATASET/RAW_IMAGES/4_Blurred/'\n",
    "path_annotation_4 = 'annotations/annotations_04_blurred.json'\n",
    "img_names_4, iou_scores_4 = calculate_iou(image_dir_4, path_annotation_4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "categories = ['1_Day_Light']*len(img_names_1) + ['2_Night']*len(img_names_2) + ['3_Tilted']*len(img_names_3) + ['4_Blurred']*len(img_names_4)\n",
    "img_names = img_names_1 + img_names_2 + img_names_3 + img_names_4\n",
    "iou_scores = iou_scores_1 + iou_scores_2 + iou_scores_3 + iou_scores_4\n",
    "dict_results = {'img_name': img_names, 'category': categories, 'iou_score': iou_scores}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "df_results = pd.DataFrame(data=dict_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>img_name</th>\n",
       "      <th>category</th>\n",
       "      <th>iou_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>001.JPG</td>\n",
       "      <td>1_Day_Light</td>\n",
       "      <td>0.895251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>002.JPG</td>\n",
       "      <td>1_Day_Light</td>\n",
       "      <td>0.960099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>003.JPG</td>\n",
       "      <td>1_Day_Light</td>\n",
       "      <td>0.892853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>004.JPG</td>\n",
       "      <td>1_Day_Light</td>\n",
       "      <td>0.841683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>005.JPG</td>\n",
       "      <td>1_Day_Light</td>\n",
       "      <td>0.911559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163</th>\n",
       "      <td>165.JPG</td>\n",
       "      <td>4_Blurred</td>\n",
       "      <td>0.937832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164</th>\n",
       "      <td>166.JPG</td>\n",
       "      <td>4_Blurred</td>\n",
       "      <td>0.937056</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165</th>\n",
       "      <td>167.JPG</td>\n",
       "      <td>4_Blurred</td>\n",
       "      <td>0.021171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>166</th>\n",
       "      <td>168.JPG</td>\n",
       "      <td>4_Blurred</td>\n",
       "      <td>0.809365</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167</th>\n",
       "      <td>169.JPG</td>\n",
       "      <td>4_Blurred</td>\n",
       "      <td>0.824914</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>168 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    img_name     category  iou_score\n",
       "0    001.JPG  1_Day_Light   0.895251\n",
       "1    002.JPG  1_Day_Light   0.960099\n",
       "2    003.JPG  1_Day_Light   0.892853\n",
       "3    004.JPG  1_Day_Light   0.841683\n",
       "4    005.JPG  1_Day_Light   0.911559\n",
       "..       ...          ...        ...\n",
       "163  165.JPG    4_Blurred   0.937832\n",
       "164  166.JPG    4_Blurred   0.937056\n",
       "165  167.JPG    4_Blurred   0.021171\n",
       "166  168.JPG    4_Blurred   0.809365\n",
       "167  169.JPG    4_Blurred   0.824914\n",
       "\n",
       "[168 rows x 3 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_results.to_csv('results.csv', header=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "latex_results = df_results.to_latex(index=False, float_format=\"%.2f\", longtable=True)\n",
    "# print(latex_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"8\" halign=\"left\">iou_score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>category</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1_Day_Light</th>\n",
       "      <td>50.0</td>\n",
       "      <td>0.899171</td>\n",
       "      <td>0.070703</td>\n",
       "      <td>0.490584</td>\n",
       "      <td>0.880680</td>\n",
       "      <td>0.908399</td>\n",
       "      <td>0.927447</td>\n",
       "      <td>0.979789</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2_Night</th>\n",
       "      <td>48.0</td>\n",
       "      <td>0.914371</td>\n",
       "      <td>0.047579</td>\n",
       "      <td>0.733776</td>\n",
       "      <td>0.904953</td>\n",
       "      <td>0.924169</td>\n",
       "      <td>0.941920</td>\n",
       "      <td>0.989232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3_Tilted</th>\n",
       "      <td>50.0</td>\n",
       "      <td>0.894636</td>\n",
       "      <td>0.098656</td>\n",
       "      <td>0.448946</td>\n",
       "      <td>0.894125</td>\n",
       "      <td>0.918802</td>\n",
       "      <td>0.940237</td>\n",
       "      <td>0.966984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4_Blurred</th>\n",
       "      <td>20.0</td>\n",
       "      <td>0.847843</td>\n",
       "      <td>0.203574</td>\n",
       "      <td>0.021171</td>\n",
       "      <td>0.857917</td>\n",
       "      <td>0.911042</td>\n",
       "      <td>0.935889</td>\n",
       "      <td>0.951105</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            iou_score                                                    \\\n",
       "                count      mean       std       min       25%       50%   \n",
       "category                                                                  \n",
       "1_Day_Light      50.0  0.899171  0.070703  0.490584  0.880680  0.908399   \n",
       "2_Night          48.0  0.914371  0.047579  0.733776  0.904953  0.924169   \n",
       "3_Tilted         50.0  0.894636  0.098656  0.448946  0.894125  0.918802   \n",
       "4_Blurred        20.0  0.847843  0.203574  0.021171  0.857917  0.911042   \n",
       "\n",
       "                                 \n",
       "                  75%       max  \n",
       "category                         \n",
       "1_Day_Light  0.927447  0.979789  \n",
       "2_Night      0.941920  0.989232  \n",
       "3_Tilted     0.940237  0.966984  \n",
       "4_Blurred    0.935889  0.951105  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# from statistics import mean\n",
    "# df_results.groupby(by='category').agg(\n",
    "#     {\n",
    "#          'iou_score':[mean, min, max],\n",
    "#     }\n",
    "# )\n",
    "\n",
    "df_summary = df_results.groupby(by='category').describe()\n",
    "df_summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print(df_summary.to_latex(index=False, float_format=\"%.2f\", longtable=True, multicolumn=True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:title={'center':'iou_score'}, xlabel='category'>"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_results.boxplot(by='category', column='iou_score', figsize=(10,6))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "02f5630d1f84b678942898632511e95e6294d5141a13222a2ec8f25d72471916"
  },
  "kernelspec": {
   "display_name": "Python 3.9.7 ('bga2')",
   "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": "3.9.7"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}