Driver_Drowsiness_Detection.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Driver Drowsiness Detection.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "__1. Install and Import Dependencies__"
      ],
      "metadata": {
        "id": "aes2TEvotfo9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install torch==1.8.1+cu111 torchvision==0.9.1+cu111 torchaudio===0.8.1 -f https://download.pytorch.org/whl/lts/1.8/torch_lts.html"
      ],
      "metadata": {
        "id": "70y46ff0tlgx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/ultralytics/yolov5"
      ],
      "metadata": {
        "id": "B_7oEJaItpiX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!cd yolov5 & pip install -r requirements.txt"
      ],
      "metadata": {
        "id": "_QhYHJnOtrmm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import torch\n",
        "from matplotlib import pyplot as plt\n",
        "import numpy as np\n",
        "import cv2"
      ],
      "metadata": {
        "id": "RKnN0HkVtu5i"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "__2. Load Model__\n"
      ],
      "metadata": {
        "id": "ajXlSwflt7QI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = torch.hub.load('ultralytics/yolov5', 'yolov5s')"
      ],
      "metadata": {
        "id": "b-WP2bfIt9qD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model"
      ],
      "metadata": {
        "id": "VRkb9NxYuAz2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "3. Make Detections with Images\n"
      ],
      "metadata": {
        "id": "WzeCZM47uEap"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "img = 'https://upload.wikimedia.org/wikipedia/commons/thumb/e/e4/Cars_in_traffic_in_Auckland%2C_New_Zealand_-_copyright-free_photo_released_to_public_domain.jpg/800px-Cars_in_traffic_in_Auckland%2C_New_Zealand_-_copyright-free_photo_released_to_public_domain.jpg'"
      ],
      "metadata": {
        "id": "nLOAuNAOuIP9"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "results = model(img)\n",
        "results.print()"
      ],
      "metadata": {
        "id": "QImhOdfcuSWe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%matplotlib inline \n",
        "plt.imshow(np.squeeze(results.render()))\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "irxjFT3huVzt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "results.render()"
      ],
      "metadata": {
        "id": "xB_-5cHDuYtz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "__4. Real Time Detections__\n"
      ],
      "metadata": {
        "id": "9AcDzvzjuhqk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cap = cv2.VideoCapture(0)\n",
        "while cap.isOpened():\n",
        "    ret, frame = cap.read()\n",
        "    \n",
        "    # Make detections \n",
        "    results = model(frame)\n",
        "    \n",
        "    cv2.imshow('YOLO', np.squeeze(results.render()))\n",
        "    \n",
        "    if cv2.waitKey(10) & 0xFF == ord('q'):\n",
        "        break\n",
        "cap.release()\n",
        "cv2.destroyAllWindows()"
      ],
      "metadata": {
        "id": "LDo9yRovuiyX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "__5. Train from scratch__\n"
      ],
      "metadata": {
        "id": "sb-yZnpdultg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import uuid   # Unique identifier\n",
        "import os\n",
        "import time"
      ],
      "metadata": {
        "id": "uXcscCY1uoFP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "IMAGES_PATH = os.path.join('data', 'images') #/data/images\n",
        "labels = ['awake', 'drowsy']\n",
        "number_imgs = 5"
      ],
      "metadata": {
        "id": "hmvbQknLuro1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cap = cv2.VideoCapture(0)\n",
        "# Loop through labels\n",
        "for label in labels:\n",
        "    print('Collecting images for {}'.format(label))\n",
        "    time.sleep(5)\n",
        "    \n",
        "    # Loop through image range\n",
        "    for img_num in range(number_imgs):\n",
        "        print('Collecting images for {}, image number {}'.format(label, img_num))\n",
        "        \n",
        "        # Webcam feed\n",
        "        ret, frame = cap.read()\n",
        "        \n",
        "        # Naming out image path\n",
        "        imgname = os.path.join(IMAGES_PATH, label+'.'+str(uuid.uuid1())+'.jpg')\n",
        "        \n",
        "        # Writes out image to file \n",
        "        cv2.imwrite(imgname, frame)\n",
        "        \n",
        "        # Render to the screen\n",
        "        cv2.imshow('Image Collection', frame)\n",
        "        \n",
        "        # 2 second delay between captures\n",
        "        time.sleep(2)\n",
        "        \n",
        "        if cv2.waitKey(10) & 0xFF == ord('q'):\n",
        "            break\n",
        "cap.release()\n",
        "cv2.destroyAllWindows()"
      ],
      "metadata": {
        "id": "F3TOSh8duuYB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(os.path.join(IMAGES_PATH, labels[0]+'.'+str(uuid.uuid1())+'.jpg'))"
      ],
      "metadata": {
        "id": "3YLlP6WVuxqY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for label in labels:\n",
        "    print('Collecting images for {}'.format(label))\n",
        "    for img_num in range(number_imgs):\n",
        "        print('Collecting images for {}, image number {}'.format(label, img_num))\n",
        "        imgname = os.path.join(IMAGES_PATH, label+'.'+str(uuid.uuid1())+'.jpg')\n",
        "        print(imgname)   "
      ],
      "metadata": {
        "id": "zzvUuyR5u1ob"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/tzutalin/labelImg"
      ],
      "metadata": {
        "id": "a5lhQS1Eu4bN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install pyqt5 lxml --upgrade\n",
        "!cd labelImg && pyrcc5 -o libs/resources.py resources.qrc"
      ],
      "metadata": {
        "id": "pFYPtjwKu7Kg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!cd yolov5 && python train.py --img 320 --batch 16 --epochs 500 --data dataset.yml --weights yolov5s.pt --workers 2"
      ],
      "metadata": {
        "id": "WZP5i44KvDBW"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}