OMR_Grader.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d7eb769a",
   "metadata": {},
   "source": [
    "**Import Necessary packages**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f3d00336",
   "metadata": {},
   "outputs": [],
   "source": [
    "from imutils.perspective import four_point_transform\n",
    "from imutils import contours\n",
    "import numpy as np\n",
    "import argparse\n",
    "import imutils\n",
    "import cv2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3728604e",
   "metadata": {},
   "source": [
    "**define the answer key which maps the question number to the correct answer**\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ece43d2f",
   "metadata": {},
   "outputs": [],
   "source": [
    "ANSWER_KEY= {0: 1,1: 4,2: 0,3: 3,4: 1}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72b0a777",
   "metadata": {},
   "source": [
    "**load the image, convert it to grayscale, blur it slightly , then find edges**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ef99fd2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "image = cv2.imread(\"test_05.png\")\n",
    "gray = cv2.cvtColor(image,cv2.COLOR_BGR2GRAY)\n",
    "blurred = cv2.GaussianBlur(gray,(5,5),0)\n",
    "edged = cv2.Canny(blurred,75,200)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55e2d808",
   "metadata": {},
   "source": [
    "**find contours in the edge map, then initialize the contour that corresponds to the document**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "afcedd29",
   "metadata": {},
   "outputs": [],
   "source": [
    "cnts = cv2.findContours(edged.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)\n",
    "cnts = imutils.grab_contours(cnts)\n",
    "docCnt = None"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9ebf20a",
   "metadata": {},
   "source": [
    "**sort the contours according to their size in descending order \n",
    "loop over the sorted contours and approximate the contour \n",
    "if our approximated contour has four points, then we can assume we have found the paper**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6df54ecc",
   "metadata": {},
   "outputs": [],
   "source": [
    "if len(cnts)>0:\n",
    "    cnts = sorted(cnts,key=cv2.contourArea,reverse=True)\n",
    "    for c in cnts:\n",
    "        peri = cv2.arcLength(c,True)\n",
    "        approx=cv2.approxPolyDP(c,0.02*peri,True)\n",
    "        if len(approx) == 4:\n",
    "            docCnt=approx\n",
    "            break"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f1aaf2a",
   "metadata": {},
   "source": [
    "**apply a four point perspective transform to both the  \n",
    "original image and grayscale image to obtain a top-down\n",
    "birds eye view of the paper**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a2bbed2a",
   "metadata": {},
   "outputs": [],
   "source": [
    "paper = four_point_transform(image,docCnt.reshape(4,2))\n",
    "warped= four_point_transform(gray,docCnt.reshape(4,2))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "85bc4914",
   "metadata": {},
   "source": [
    "**apply Otsu's thresholding method to binarize the warped piece of paper**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58222d63",
   "metadata": {},
   "outputs": [],
   "source": [
    "(T,threshInv)=cv2.threshold(warped,0,255,cv2.THRESH_BINARY_INV|cv2.THRESH_OTSU)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0fda0bea",
   "metadata": {},
   "source": [
    "**find contours in the thresholded image, then initialize\n",
    "the list of contours that correspond to questions**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a250e33",
   "metadata": {},
   "outputs": [],
   "source": [
    "cnts = cv2.findContours(threshInv.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)\n",
    "cnts = imutils.grab_contours(cnts)\n",
    "questionCnts = []"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d1f30d50",
   "metadata": {},
   "source": [
    "**compute the bounding box of the contour, then use the bounding box to derive the aspect ratio**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12534fcd",
   "metadata": {},
   "source": [
    "**in order to label the contour as a question, region should be sufficiently wide, sufficiently tall, and\n",
    "have an aspect ratio approximately equal to 1**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "89665710",
   "metadata": {},
   "outputs": [],
   "source": [
    "for c in cnts:\n",
    "    (x,y,w,h) = cv2.boundingRect(c)\n",
    "    ar = w / float(h)\n",
    "    if w>=20 and h>=20 and ar>=0.9 and ar<=1.1:\n",
    "        questionCnts.append(c)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b7869ad1",
   "metadata": {},
   "source": [
    "**sort the question contours top-to-bottom, then initialize the total number of correct answers**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "045dbbcf",
   "metadata": {},
   "outputs": [],
   "source": [
    "questionCnts = contours.sort_contours(questionCnts,method=\"top-to-bottom\")[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "75fbd49c",
   "metadata": {},
   "outputs": [],
   "source": [
    "correct=0"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79d7e040",
   "metadata": {},
   "source": [
    "**sort the contours for the current question from left to right, then initialize the index of the bubbled answer**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7db8c6c",
   "metadata": {},
   "source": [
    "**construct a mask that reveals only the current \"bubble\" for the question**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "443bf5db",
   "metadata": {},
   "source": [
    "**apply the mask to the thresholded image, then count the number of non-zero pixels in the bubble area**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e657152c",
   "metadata": {},
   "source": [
    "**if the current total has a larger number of total non-zero pixels, then we are examining the currently bubbled-in answer**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "58d9b7f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "for (q,i) in enumerate(np.arange(0,len(questionCnts),5)):\n",
    "    cnts=contours.sort_contours(questionCnts[i: i+5])[0]\n",
    "    bubbled=None\n",
    "    for(j,c) in enumerate(cnts):\n",
    "        mask=np.zeros(threshInv.shape,dtype=\"uint8\")\n",
    "        cv2.drawContours(mask,[c],-1,255,-1)\n",
    "        \n",
    "        mask=cv2.bitwise_and(threshInv,threshInv,mask=mask)\n",
    "        total=cv2.countNonZero(mask)\n",
    "        if bubbled is None or total>bubbled[0]:\n",
    "            bubbled=(total,j)\n",
    "            \n",
    "    color=(0,0,255)\n",
    "    k=ANSWER_KEY[q]\n",
    "    \n",
    "    if k==bubbled[1]:\n",
    "        color=(0,255,0)\n",
    "        correct+=1\n",
    "        \n",
    "    cv2.drawContours(paper,[cnts[k]],-1,color,3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f5b8d0a0",
   "metadata": {},
   "source": [
    "**grab the test taker**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4379c9d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "score = (correct / 5.0) * 100\n",
    "print(\"[INFO] score: {:.2f}%\".format(score))\n",
    "cv2.putText(paper, \"{:.2f}%\".format(score), (10, 30),\n",
    "    cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 0, 255), 2)\n",
    "cv2.imshow(\"Original\", image)\n",
    "cv2.imshow(\"Exam\", paper)\n",
    "cv2.waitKey(0)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}