lecture-3 Edge detection.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T08:41:54.473505Z",
     "start_time": "2020-04-29T08:41:53.489984Z"
    },
    "code_folding": [
     7
    ]
   },
   "outputs": [],
   "source": [
    "Image = \"/mnt/Study_Storage/Documents/Datasets/Images/standard_test_images/lena_gray_256.tif\" #gray\n",
    "\n",
    "import cv2 \n",
    "import skimage  \n",
    "import numpy as np \n",
    "\n",
    "\n",
    "def show(img):\n",
    "    cv2.imshow(\"image\", img)\n",
    "    cv2.waitKey(0)\n",
    "    cv2.destroyAllWindows()\n",
    "    \n",
    "\n",
    "img = cv2.imread(Image, 0)\n",
    "show(img)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "### precision n recall"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {
    "hidden": true
   },
   "source": [
    " - Precision and recall \n",
    " - Four things: \n",
    "         - if we predict Positive: it can be true positive or false position, that is:\n",
    "                 - true positive: intersection b/w correct and prediction\n",
    "                 - false positive: we predicted positive but it's not positive\n",
    "          - if we predict Negative that can be true negative or false negative\n",
    "           that is:\n",
    "                   - true negative: it's correct prediction that it's negative\n",
    "                   - false negative: we predict it's negative but it's not negative in real.\n",
    "\n",
    "\n",
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "### prewit and sobel and scharr\n",
    "\n",
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "hidden": true
   },
   "source": [
    "```python \n",
    "\n",
    "prewit < sobel < scharr in order of rotation invariance \n",
    "\n",
    "Return the square root of the sum of squares of the horizontal\n",
    "and vertical Prewitt transforms. The edge magnitude depends slightly\n",
    "on edge directions, since the approximation of the gradient operator by\n",
    "the Prewitt operator is not completely rotation invariant. For a better\n",
    "rotation invariance, the Scharr operator should be used. The Sobel operator\n",
    "has a better rotation invariance than the Prewitt operator, but a worse\n",
    "rotation invariance than the Scharr operator. \n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T09:16:22.106835Z",
     "start_time": "2020-04-29T09:16:13.829717Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "img_prewit_x = filters.edges.prewitt_h(img)\n",
    "img_prewit_y = filters.edges.prewitt_v(img)\n",
    "img_prewit = filters.edges.prewitt(img) \n",
    "\n",
    "\n",
    "'''Return the square root of the sum of squares of the horizontal\n",
    "and vertical Prewitt transforms. The edge magnitude depends slightly\n",
    "on edge directions, since the approximation of the gradient operator by\n",
    "the Prewitt operator is not completely rotation invariant. For a better\n",
    "rotation invariance, the Scharr operator should be used. The Sobel operator\n",
    "has a better rotation invariance than the Prewitt operator, but a worse\n",
    "rotation invariance than the Scharr operator.'''\n",
    "\n",
    "show(img_prewit_x)\n",
    "show(img_prewit_y)\n",
    "show(img_prewit)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T09:19:12.413270Z",
     "start_time": "2020-04-29T09:19:05.497717Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "img_sobel_x = filters.edges.sobel_h(img)\n",
    "img_sobel_y = filters.edges.sobel_v(img)\n",
    "img_sobel = filters.edges.sobel(img)\n",
    "\n",
    "show(img_sobel_x)\n",
    "show(img_sobel_y)\n",
    "show(img_sobel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T09:20:44.225386Z",
     "start_time": "2020-04-29T09:20:36.740787Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "img_scharr_x = filters.edges.scharr_h(img)\n",
    "img_scharr_y = filters.edges.scharr_v(img)\n",
    "img_scharr = filters.edges.scharr(img)\n",
    "\n",
    "show(img_scharr_x)\n",
    "show(img_scharr_y)\n",
    "show(img_scharr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "### decomposition of 2 d into 2 1d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T09:55:28.033883Z",
     "start_time": "2020-04-29T09:55:13.645912Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 3)\n",
      "(3, 1)\n"
     ]
    }
   ],
   "source": [
    "one_d = np.array([[1,1,1]])/3 \n",
    "\n",
    "print(one_d.shape)\n",
    "\n",
    "img_1_d = filters.edges.convolve(img, weights= one_d)\n",
    "show(img_1_d)\n",
    "\n",
    "one_d = np.array([[1],[1],[1]])/3 \n",
    "print(one_d.shape)\n",
    "show(filters.edges.convolve(img, weights= one_d)) \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T09:54:22.236505Z",
     "start_time": "2020-04-29T09:54:22.233581Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3, 1)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "heading_collapsed": true
   },
   "source": [
    "### laplace based"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T10:02:04.470515Z",
     "start_time": "2020-04-29T10:02:00.450530Z"
    },
    "hidden": true
   },
   "outputs": [],
   "source": [
    "''' laplacian '''\n",
    "from skimage import filters \n",
    "\n",
    "img_smooth = filters.gaussian(img, sigma= 1)\n",
    "img_laplace = filters.laplace(img_smooth) \n",
    "\n",
    "# find zero crossings that is not done here \n",
    "\n",
    "# that it will be marr-hildreath edge detector \n",
    "\n",
    "\n",
    "show(img_laplace)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T10:04:50.788832Z",
     "start_time": "2020-04-29T10:04:50.776329Z"
    },
    "hidden": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(17949,) (17949,) (256, 256)\n"
     ]
    }
   ],
   "source": [
    "zero_crossings = np.where(np.diff(np.sign(img_laplace)))\n",
    "print(zero_crossings[0].shape, zero_crossings[1].shape, img_laplace.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### canny"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T08:59:41.772762Z",
     "start_time": "2020-04-29T08:59:41.762385Z"
    }
   },
   "outputs": [],
   "source": [
    "from skimage import feature "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T10:39:48.220149Z",
     "start_time": "2020-04-29T10:39:38.758385Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "canny = feature.canny(img, sigma = 1.0, low_threshold= 15, high_threshold= 35) # mask of binary edge points\n",
    "canny_edge = np.multiply(canny, img)\n",
    "show(canny_edge)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-04-29T10:41:42.289608Z",
     "start_time": "2020-04-29T10:41:42.255835Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Search results for 'laplace'\n",
      "----------------------------\n",
      "skimage.filters.laplace\n",
      "    Find the edges of an image using the Laplace operator.\n",
      "skimage.filters.tests.test_edges.test_laplace_mask\n",
      "    Laplace on a masked array should be zero.\n",
      "skimage.filters.tests.test_edges.test_laplace_zeros\n",
      "    Laplace on a square image.\n"
     ]
    }
   ],
   "source": [
    "skimage.lookfor(\"laplace\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "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": "3.7.7"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}