01-Read-Prepare-Save-Data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "659e3e1f",
   "metadata": {},
   "source": [
    "## Import necessary libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "ef76b759",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from glob import glob\n",
    "from tqdm.notebook import tqdm\n",
    "import pickle\n",
    "\n",
    "import cv2\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1d493081",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "b094673f",
   "metadata": {},
   "source": [
    "## Define path of prepared data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "222d609d",
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = [\"benign\", \"malignant\", \"normal\"]\n",
    "\n",
    "filenames_template = \"data/Dataset_BUSI/Dataset_BUSI_with_GT/{}/{} ({})\"\n",
    "class_template = \"data/Dataset_BUSI/Dataset_BUSI_with_GT/{}\"\n",
    "\n",
    "saved_directory = \"prepared_data\"\n",
    "saved_filename = \"image_with_mask_label.pickle\"\n",
    "\n",
    "os.makedirs(saved_directory, exist_ok=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "09b938a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "total_image = 0\n",
    "filename_dict = {}\n",
    "for classe in classes:\n",
    "    filename_dict[classe] = {}\n",
    "\n",
    "for classe in classes:\n",
    "    flag = True\n",
    "    counter = 1\n",
    "    \n",
    "    while flag:\n",
    "        image_filename = filenames_template.format(classe, classe, counter)\n",
    "        \n",
    "        #print(image_filename, os.path.exists(image_filename + \".png\"))\n",
    "        if (not os.path.exists(image_filename + \".png\")):\n",
    "            break\n",
    "            \n",
    "        mask_filenames = glob(image_filename + \"_*.png\")\n",
    "        \n",
    "        filename_dict[classe][counter] = [image_filename + \".png\", mask_filenames]\n",
    "        \n",
    "        counter += 1\n",
    "        total_image += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ae3d90c5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "20c52814",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "benign\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ea2863f1072a4e6698842dcdf5c92e08",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/437 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "malignant\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "95fdd43ff28a4155a53d29554fbd318c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/210 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "normal\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a109f114faf24de48a1ed3183ddea06f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/133 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = [] \n",
    "image_size = (128, 128)\n",
    "\n",
    "N_labels = {}\n",
    "\n",
    "for classe in classes:\n",
    "    print(classe)\n",
    "    N_labels[classe] = len(filename_dict[classe].keys())\n",
    "    for counter in tqdm(range(1, N_labels[classe] + 1)):\n",
    "        filenames = filename_dict[classe][counter]\n",
    "        \n",
    "        image = cv2.imread(filenames[0])\n",
    "        image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)\n",
    "        image = cv2.resize(image, image_size)\n",
    "        \n",
    "        mask = np.zeros(image_size)\n",
    "        for filename in filenames[1]:\n",
    "            buff = cv2.imread(filename)\n",
    "            buff = cv2.cvtColor(buff, cv2.COLOR_BGR2GRAY)\n",
    "            buff = cv2.resize(buff, image_size)\n",
    "            \n",
    "            mask += buff\n",
    "        mask = (mask > 0).astype(np.int8)\n",
    "        \n",
    "        \n",
    "        data.append({\"image\": image, \"mask\": mask, \"label\": classes.index(classe)})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "bc0fbc4f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fed2af143a0>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "## show a image\n",
    "index = 437\n",
    "\n",
    "plt.figure()\n",
    "plt.imshow(data[index][\"image\"] * data[index][\"mask\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "651b3c52",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'benign': 437, 'malignant': 210, 'normal': 133}"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "N_labels # label with the number of data in each class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d04665b6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "82917965",
   "metadata": {},
   "source": [
    "## Save processed image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "1a3f6e4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(os.path.join(saved_directory, saved_filename), 'wb') as handle:\n",
    "    pickle.dump(data, handle, protocol=pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "4089435f",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(os.path.join(saved_directory, saved_filename), 'rb') as handle:\n",
    "    buffer = pickle.load(handle)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "68b37606",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "780"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(buffer) # total number of image"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.8.8 64-bit ('TensorFlow2': conda)",
   "language": "python",
   "name": "python388jvsc74a57bd0e32a148bad2c5c70e476b03c767e5d0c85b32db8213e03ce69283c8b3ecdedb5"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}