Collecting Images.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I created virtual environment that is called \"first\" in my laptop. \n",
    "And activated it. \n",
    "Finally, added tvirtual environment to the Python Kernel.\n",
    "\n",
    "code: \n",
    "    python -m venv tfod\n",
    "    \n",
    "    source tfod/bin/activate # Linux\n",
    "    .\\tfod\\Scripts\\activate # Windows \n",
    "    \n",
    "    python -m pip install --upgrade pip\n",
    "    pip install ipykernel\n",
    "    python -m ipykernel install --user --name=firstj\n",
    "    \n",
    "And, opened jupyter notebook in firstj kernel.!!\n",
    "\n",
    "\n",
    
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Import Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install opencv\n",
    "!pip install uuid \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import opencv\n",
    "import cv2 \n",
    "\n",
    "# Import uuid\n",
    "#I can name my images uniquely. \n",
    "import uuid\n",
    "\n",
    "import os\n",
    "\n",
    "# for using some sleep function\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Define Images to Collect"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = ['vletter', 'thumbsdown', 'thumbsup', 'call' ]\n",
    "collect_imgs = 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I'll collect 5 images for each one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Setup Folders "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "IMAGES_PATH = os.path.join('Tensorflow', 'workspace', 'images', 'collectedimages')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "IMAGES_PATH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.name\n",
    "#posix for linux but nt for windows."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not os.path.exists(IMAGES_PATH):\n",
    "    if os.name == 'posix':\n",
    "        !mkdir -p {IMAGES_PATH}\n",
    "    if os.name == 'nt':\n",
    "         !mkdir {IMAGES_PATH}\n",
    "for label in labels:\n",
    "    path = os.path.join(IMAGES_PATH, label)\n",
    "    if not os.path.exists(path):\n",
    "        !mkdir {path}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we can see 6 folders that are for vletter, phone, thumbsup etc. (on the path of course)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.chdir(IMAGES_PATH)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "cwd = os.getcwd()\n",
    "cwd\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. Capture Images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Videocapture will connect with our webcam. \"0\" this number is important, maybe we may have to change it (like 1,2 etc.)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for label in labels:\n",
    "    cap = cv2.VideoCapture(\"/dev/video0\")\n",
    "    print('Collecting images for {}'.format(label))\n",
    "    time.sleep(5)\n",
    "    for imgnum in range(collect_imgs):\n",
    "        print('Collecting image {}'.format(imgnum))\n",
    "        ret, frame = cap.read()\n",
    "        imgname = os.path.join(cwd,label,label+'.'+'{}.jpg'.format(str(uuid.uuid1())))        \n",
    "        cv2.imwrite(imgname, frame)\n",
    "        cv2.imshow('frame', frame)\n",
    "        time.sleep(2)\n",
    "        \n",
    "    \n",
    "        \n",
    "        if cv2.waitKey(1) & 0xFF == ord('q'):\n",
    "            break\n",
    "    cap.release() # RELEASE VIDEO HERE\n",
    "    \n",
    "cap.release() # NOT HERE!!!\n",
    "cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 5. Image Labelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip uninstall --upgrade pyqt5 lxml"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#install pyqt5 and lxml library."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "LABEL_PATH = os.path.join('Tensorflow', 'labelimg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.chdir(\"{LABELIMG_PATH}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not os.path.exists(cwd):\n",
    "    os.chdir(\"{LABELIMG_PATH}\")\n",
    "#   !git clone https://github.com/tzutalin/labelImg {LABELIMG_PATH}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if os.name == 'posix':\n",
    "    !python labelimg.py\n",
    "if os.name =='nt':\n",
    "    !cd {LABELIMG_PATH} && pyrcc5 -o libs/resources.py resources.qrc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "cwd = os.getcwd()\n",
    "cwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!python labelimg.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Then, use this app for labeling. \n",
    "1. Open your images folder\n",
    "2. Press \"W\" button and draw a frame. \n",
    "3. Give true name for your each image. (Eg: ThumbsUp)\n",
    "4. Ctrl+S >> Save it. \n",
    "Do these for each label."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 6. Move them into a Training and Testing Partition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAIN_PATH = os.path.join('Tensorflow', 'workspace', 'images', 'train')\n",
    "TEST_PATH = os.path.join('Tensorflow', 'workspace', 'images', 'test')\n",
    "ARCHIVE_PATH = os.path.join('Tensorflow', 'workspace', 'images', 'archive.tar.gz')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!tar -czf {ARCHIVE_PATH} {TRAIN_PATH} {TEST_PATH}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All labeled images and images can be copied to train and test folders. \n" , 
     "\n",

    "I compiled my code with @nicknochnack.\n",
      "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tfs-venv",
   "language": "python",
   "name": "tfs-venv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}