ps4.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## you can take an analytic derivative of a Gaussian in X or Y and use that filter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "simA = cv2.imread('input/simA.jpg', cv2.IMREAD_GRAYSCALE)\n",
    "plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculateGradients(img, ksize, axis=0):\n",
    "    if axis == 1:\n",
    "        grad = cv2.Sobel(img, cv2.CV_64F, 0, 1, ksize)\n",
    "    else:\n",
    "        grad = cv2.Sobel(img, cv2.CV_64F, 1, 0, ksize)\n",
    "    \n",
    "    grad = cv2.normalize(grad, grad, alpha=0, beta=255, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_8U)\n",
    "    return grad"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "gradx = calculateGradients(simA, 3, axis=1)\n",
    "plt.imshow(gradx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def harrisScoreFunction(Ix, Iy, weights, alpha, norm=True):\n",
    "    Ixx = Ix ** 2\n",
    "    Ixy = Ix * Iy\n",
    "    Iyx = Iy * Ix\n",
    "    Iyy = Iy ** 2\n",
    "    \n",
    "    im_rows, im_cols = Ix.shape\n",
    "    w_rows, w_cols = weights.shape\n",
    "    \n",
    "    Rs = np.zeros(Ix.shape)\n",
    "    \n",
    "    for i in range(w_rows / 2, im_rows - w_rows / 2):\n",
    "        min_i = i - w_rows / 2\n",
    "        max_i = i + w_rows / 2 + 1\n",
    "        for j in range(w_cols / 2, im_cols - im_cols / 2):\n",
    "            min_j = j - w_cols / 2\n",
    "            max_j = j + w_cols / 2 + 1\n",
    "            \n",
    "            wIxx = Ixx[min_i:max_i, min_j:max_j]\n",
    "            wIxy = Ixy[min_i:max_i, min_j:max_j]\n",
    "            wIyx = Iyx[min_i:max_i, min_j:max_j]\n",
    "            wIyy = Iyy[min_i:max_i, min_j:max_j]\n",
    "            \n",
    "            Mxx = (weights * wIxx).sum()\n",
    "            Mxy = (weights * wIxy).sum()\n",
    "            Myx = (weights * wIyx).sum()\n",
    "            Myy = (weights * wIyy).sum()\n",
    "            \n",
    "            M = np.asarray([Mxx, Mxy, Myx, Myy]).reshape((2,2))\n",
    "            Rs[i, j] = np.linalg.det(M) - alpha * (M.trace() ** 2)\n",
    "    if norm:\n",
    "        Rs = cv2.normalize(Rs, Rs, alpha=0, beta=255,\n",
    "                           norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_8U)\n",
    "    return Rs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generateGaussianWeights(window_size):\n",
    "    c = np.zeros((window_size,)*2, dtype=np.float32);\n",
    "    c[window_size / 2, window_size / 2] = 1.0\n",
    "    w = cv2.GaussianBlur(c, (window_size,)*2, 0)\n",
    "    return w"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "gradx = calculateGradients(simA, 3, axis=1)\n",
    "grady = calculateGradients(simA, 3, axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "weights = generateGaussianWeights(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Rs = harrisScoreFunction(grad_x, grad_y, weights, 0.04)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.imshow(Rs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def findCorner(scores, threshold, radius):\n",
    "    shape = scores.shape\n",
    "    scores[scores < threshold] = 0\n",
    "    scores = scores.reshape(shape)\n",
    "    \n",
    "    #Non-maximum suppression\n",
    "    for i in range(radius, shape[0] - radius):\n",
    "        min_i = i - radius\n",
    "        max_i = i + radius\n",
    "        for j in range(radius, shape[1] - radius):\n",
    "            min_j = j - radius\n",
    "            max_j = j + radius\n",
    "            \n",
    "            patch = scores[min_i:max_i, min_j:max_j]\n",
    "            max_val = np.amax(patch)\n",
    "            patch[patch < max_val] = 0\n",
    "            scores[min_i:max_i, min_j:max_j] = patch\n",
    "    \n",
    "    corners = []\n",
    "    for i in range(shape[0]):\n",
    "        for j in range(shape[1]):\n",
    "            if scores[i, j] > 0:\n",
    "                corners.append((i, j))\n",
    "    return corners"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "corners = findCorner(Rs, 100, 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "imgA_color = cv2.imread('input/simA.jpg')\n",
    "for r,c in corners:\n",
    "    cv2.circle(imgA_color, (c,r), 2, (0,0,255))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "imgA_color = cv2.cvtColor(imgA_color, cv2.COLOR_BGR2RGB)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.imshow(imgA_color)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}