{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A notebook to test calculating of SNR, SSIM, and MSE for a filter using a number of randomly generated additive noise problems"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Setup\n",
    "Here we import the libraries we need"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt # plotting and showing images\n",
    "import numpy as np # handling arrays\n",
    "from skimage.io import imread # reading images\n",
    "#from skimage.measure import compare_ssim as ssim \n",
    "from skimage.metrics import structural_similarity as ssim # structural similarity\n",
    "mse = lambda img1, img2: np.sum(np.power(img1-img2,2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load and Show Phantom"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "d=np.mean(imread('data/testpattern.png'),2)\n",
    "plt.imshow(d, cmap= 'gray')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tasks\n",
    "Evaluate a number of different filters using different noise levels and determine which works the best under which circumstances. The example is done for a uniform filter, and you job is to add \n",
    "\n",
    " 1. gaussian\n",
    " 2. median  \n",
    " 3. anisotropic diffusion (copy code from exercise 3) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Filter evaluation using uniform filter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "scales = [0.1, 0.5, 1, 10, 20, 100];\n",
    "\n",
    "Ntests = 10;\n",
    "\n",
    "# initialize arrays for results\n",
    "mse1 = np.zeros((len(scales), Ntests), dtype = np.float32)\n",
    "ssim1 = np.zeros((len(scales), Ntests), dtype = np.float32)\n",
    "from scipy.ndimage.filters import uniform_filter\n",
    "current_filter = lambda img: uniform_filter(img, 5)\n",
    "\n",
    "for i, c_scale in enumerate(scales):\n",
    "    for j in range(Ntests):\n",
    "        x = current_filter(d+c_scale*np.random.uniform(-c_scale, c_scale, size = d.shape))\n",
    "        mse1[i,j]=mse(d,x);\n",
    "        ssim1[i,j]=ssim(d,x);    \n",
    "    # Add some lines here to display the latest image in a subplot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(1,2)\n",
    "ax1.loglog(scales,np.mean(mse1,1)) # Add annotations for the plot and axes\n",
    "ax1.set_title('MSE vs Scale')\n",
    "ax1.set_xlabel('Scale')\n",
    "ax1.set_ylabel('MSE')\n",
    "\n",
    "ax2.semilogx(scales,np.mean(ssim1,1))\n",
    "ax2.set_title('SSIM vs Scale')\n",
    "ax2.set_xlabel('Scale')\n",
    "ax2.set_ylabel('SSIM')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make a version of the plot showing SNR instead of scale"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test run for Gauss filter\n",
    "\n",
    "### repeat the code from filter 1 \n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test run for Median filter\n",
    "\n",
    "### repeat the code from filter 1 \n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test run for Diffusion filter\n",
    "\n",
    "### repeat the code from filter 1 \n",
    "\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summarize all results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}