2021CSB1102_Kartik_PA2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "ModuleNotFoundError",
     "evalue": "No module named 'imutils'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
      "\u001b[1;32mc:\\Users\\Kartik\\OneDrive\\Desktop\\cs517\\pa2\\2021CSB1102_Kartik_PA2.ipynb Cell 1\u001b[0m in \u001b[0;36m<cell line: 8>\u001b[1;34m()\u001b[0m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=0'>1</a>\u001b[0m \u001b[39m'''\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=1'>2</a>\u001b[0m \u001b[39mKartik Tiwari\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=2'>3</a>\u001b[0m \u001b[39m- Face Morphing program that takes two images as input and generates n frams (specified by user) to perform face morphing \u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=3'>4</a>\u001b[0m \u001b[39mUsing delaunay triangulation. Please ensure that the dimensions of both images in input are same.\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=4'>5</a>\u001b[0m \u001b[39m\"shape_predictor_68_face_landmarks.dat\" used to detect 68 landmark points on face\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=5'>6</a>\u001b[0m \u001b[39m'''\u001b[39;00m\n\u001b[1;32m----> <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=7'>8</a>\u001b[0m \u001b[39mfrom\u001b[39;00m \u001b[39mimutils\u001b[39;00m \u001b[39mimport\u001b[39;00m face_utils  \u001b[39m#For face points mapping\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=8'>9</a>\u001b[0m \u001b[39mimport\u001b[39;00m \u001b[39mdlib\u001b[39;00m                     \u001b[39m#For the model to detect landmarks automatically\u001b[39;00m\n\u001b[0;32m     <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=9'>10</a>\u001b[0m \u001b[39m# https://www.geeksforgeeks.org/how-to-install-dlib-library-for-python-in-windows-10/\u001b[39;00m\n\u001b[0;32m     <a href='vscode-notebook-cell:/c%3A/Users/Kartik/OneDrive/Desktop/cs517/pa2/2021CSB1102_Kartik_PA2.ipynb#W0sZmlsZQ%3D%3D?line=10'>11</a>\u001b[0m \u001b[39m# To install dlib, first pip install cmake, then pip install dlib\u001b[39;00m\n",
      "\u001b[1;31mModuleNotFoundError\u001b[0m: No module named 'imutils'"
     ]
    }
   ],
   "source": [
    "'''\n",
    "Kartik Tiwari\n",
    "- Face Morphing program that takes two images as input and generates n frams (specified by user) to perform face morphing \n",
    "Using delaunay triangulation. Please ensure that the dimensions of both images in input are same.\n",
    "\"shape_predictor_68_face_landmarks.dat\" used to detect 68 landmark points on face\n",
    "'''\n",
    "\n",
    "from imutils import face_utils  #For face points mapping\n",
    "import dlib                     #For the model to detect landmarks automatically\n",
    "# https://www.geeksforgeeks.org/how-to-install-dlib-library-for-python-in-windows-10/\n",
    "# To install dlib, first pip install cmake, then pip install dlib\n",
    "import os\n",
    "import cv2                      #OpenCV library\n",
    "import numpy             #Numpy library\n",
    "import imageio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#To return indices of triangles in image\n",
    "def find_i(point, tr):\n",
    "    cnt = 0\n",
    "    for index in tr:\n",
    "        if point == index:\n",
    "            return cnt\n",
    "        cnt += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Delaunay triangulations\n",
    "def triangulate(img, point):\n",
    "    #Subdiv class instance\n",
    "    subs = cv2.Subdiv2D((0, 0, img.shape[1], img.shape[0]))\n",
    "    for p in point:\n",
    "        subs.insert(p)\n",
    "    #Returns triangles vertices\n",
    "    return subs.getTriangleList()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Combine input and output via affine transform\n",
    "def combineResult(img1, img2, img, t1, t2, t, alpha):\n",
    "    rect1 = cv2.boundingRect(numpy.float32([t1]))\n",
    "    rect2 = cv2.boundingRect(numpy.float32([t2]))\n",
    "    rect = cv2.boundingRect(numpy.float32([t]))\n",
    "    t1_cropped, t2_cropped, t_cropped = [],[],[]\n",
    "\n",
    "    for i in range(3):\n",
    "        t1_cropped.append(((t1[i][0] - rect1[0]), (t1[i][1] - rect1[1])))\n",
    "        t2_cropped.append(((t2[i][0] - rect2[0]), (t2[i][1] - rect2[1])))\n",
    "        t_cropped.append(((t[i][0] - rect[0]), (t[i][1] - rect[1])))\n",
    "\n",
    "    mask = numpy.zeros((rect[3], rect[2], 3), dtype=numpy.float32)\n",
    "    cv2.fillPoly(mask, [numpy.int32(t_cropped)], (1.0, 1.0, 1.0), 16)\n",
    "\n",
    "    img1_roi = img1[rect1[1]:rect1[1] + rect1[3], rect1[0]:rect1[0] + rect1[2]]\n",
    "    img2_roi = img2[rect2[1]:rect2[1] + rect2[3], rect2[0]:rect2[0] + rect2[2]]\n",
    "\n",
    "    size = (rect[2], rect[3])\n",
    "    warp1 = cv2.getAffineTransform(numpy.float32(t1_cropped), numpy.float32(t_cropped))\n",
    "    warp_img1 = cv2.warpAffine(img1_roi, warp1, size)\n",
    "    warp2 = cv2.getAffineTransform(numpy.float32(t2_cropped), numpy.float32(t_cropped))\n",
    "    warp_img2 = cv2.warpAffine(img2_roi, warp2, size)\n",
    "\n",
    "    result_roi = (1. - alpha) * warp_img1 + alpha * warp_img2\n",
    "\n",
    "    img[rect[1]:rect[1]+rect[3], rect[0]:rect[0]+rect[2]] = img[rect[1]:rect[1]+rect[3], rect[0]:rect[0]+rect[2]] * (1-mask) + result_roi * mask\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Mark the landmarks and return coordinates\n",
    "def get_points(img):\n",
    "    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
    "    rects = detector(gray, 0)\n",
    "    for (i, rect) in enumerate(rects):\n",
    "        shaped = predictor(gray, rect)\n",
    "        shaped = face_utils.shape_to_np(shaped)\n",
    "    h, w, c = img.shape\n",
    "    triangle = []\n",
    "    for coord in shaped:\n",
    "        triangle.append((int(coord[0]),int(coord[1])))\n",
    "    triangle.extend([(0,0), (0,h-1), (w-1,h-1), (w-1,0)])\n",
    "    return triangle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if __name__ == '__main__':\n",
    "    frames = int(input(\"How many frames to generate?\")) #20 recommended (optimal)\n",
    "    fps = frames/2\n",
    "    #Open the model to detect face\n",
    "    p = \"shape_predictor_68_face_landmarks.dat\"\n",
    "    detector = dlib.get_frontal_face_detector()\n",
    "    predictor = dlib.shape_predictor(p)\n",
    "\n",
    "    #Path of the two images to be morphed\n",
    "    image = cv2.imread('a1.jpg')\n",
    "    image1 = cv2.imread('b1.jpg')\n",
    "\n",
    "    triangle1 = []\n",
    "    triangle2 = []\n",
    "    print(\"Get tiepoint manually (press 1) or automatically (press 2)\")\n",
    "    n = int(input())\n",
    "    if(n==1):\n",
    "        with open(\"file1.txt\", \"r\") as f:\n",
    "            for i in f:\n",
    "                a, b, c, d = map(int, i.split())\n",
    "                triangle1.append((a, b))\n",
    "                triangle2.append((c, d))\n",
    "\n",
    "    if(n==2):\n",
    "        triangle1 = get_points(image)\n",
    "        triangle2 = get_points(image1)\n",
    "\n",
    "    dlny1 = triangulate(image, triangle1)\n",
    "\n",
    "    triangle_index = []\n",
    "    for t in dlny1:\n",
    "        pt1, pt2, pt3 = (t[0], t[1]), (t[2], t[3]), (t[4], t[5])\n",
    "        # Creating a combined point of (x,y)\n",
    "        add = [(find_i(pt1, triangle1), find_i(pt2, triangle1), find_i(pt3, triangle1))]\n",
    "        triangle_index.extend(add)\n",
    "\n",
    "    #Create a video from the generated frames\n",
    "    fourcc = cv2.VideoWriter_fourcc(*\"mp4v\")\n",
    "    h, w, c = image.shape\n",
    "    \n",
    "    fourcc = cv2.VideoWriter_fourcc(*'mp4v')\n",
    "    height, width, channels = image.shape\n",
    "    fps = frames\n",
    "\n",
    "    parent_dir = './'\n",
    "    v_dir = 'output_gif'\n",
    "    img_dir = 'output_images'\n",
    "    path = os.path.join(parent_dir, v_dir)\n",
    "    os.mkdir(path)\n",
    "    path = os.path.join(parent_dir, img_dir)\n",
    "    os.mkdir(path)\n",
    "\n",
    "    output_path = './output_images'\n",
    "    gif_path = './output_gif/morph.gif'\n",
    "\n",
    "    with imageio.get_writer(gif_path, mode='I', duration=0.05) as writer:\n",
    "        for frame in range(frames):\n",
    "            alpha_factor = (frame + 1) / frames\n",
    "            triangle_middle = []\n",
    "            \n",
    "            for i in range(len(triangle1)):\n",
    "                x = int(((1-alpha_factor) * triangle1[i][0]) + (alpha_factor * triangle2[i][0]))\n",
    "                y = int(((1-alpha_factor) * triangle1[i][1]) + (alpha_factor * triangle2[i][1]))\n",
    "                triangle_middle.append((x, y))\n",
    "            \n",
    "            morphed_image = numpy.zeros(image.shape, dtype=image.dtype)\n",
    "            \n",
    "            for j in range(len(triangle_index)):\n",
    "                x, y, z = triangle_index[j][0], triangle_index[j][1], triangle_index[j][2]\n",
    "                t1 = [triangle1[x], triangle1[y], triangle1[z]]\n",
    "                t2 = [triangle2[x], triangle2[y], triangle2[z]]\n",
    "                t = [triangle_middle[x], triangle_middle[y], triangle_middle[z]]\n",
    "                combineResult(image, image1, morphed_image, t1, t2, t, alpha_factor)\n",
    "            \n",
    "            cv2.imshow('Morphed Face', numpy.uint8(morphed_image))\n",
    "            cv2.imwrite(f'{output_path}/{frame}.jpg', morphed_image)\n",
    "            \n",
    "            gif_image = cv2.cvtColor(morphed_image, cv2.COLOR_BGR2RGB)\n",
    "            writer.append_data(gif_image)\n",
    "            cv2.waitKey(50)\n",
    "\n",
    "    cv2.destroyAllWindows()"
   ]
  }
 ],
 "metadata": {
  "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": "3.10.2"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "3cd8c85a19c845fb3699658d1408020d6f27505abef460e0c26eca2a4ff43300"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}