W2d4

week-2/solo-challenges/data-augmentation.ipynb

@@ -0,0 +1,205 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {
+    "toc": true
+   },
+   "source": [
+    "<h1>Table of Contents<span class=\"tocSkip\"></span></h1>\n",
+    "<div class=\"toc\"><ul class=\"toc-item\"><li><span><a href=\"#Data-Augmentation\" data-toc-modified-id=\"Data-Augmentation-1\"><span class=\"toc-item-num\">1&nbsp;&nbsp;</span>Data Augmentation</a></span><ul class=\"toc-item\"><li><span><a href=\"#Code-Challenge:-Importing-Images\" data-toc-modified-id=\"Code-Challenge:-Importing-Images-1.1\"><span class=\"toc-item-num\">1.1&nbsp;&nbsp;</span>Code Challenge: Importing Images</a></span></li><li><span><a href=\"#Code-Challenge:-random_rotation\" data-toc-modified-id=\"Code-Challenge:-random_rotation-1.2\"><span class=\"toc-item-num\">1.2&nbsp;&nbsp;</span>Code Challenge: <code>random_rotation</code></a></span></li><li><span><a href=\"#Code-Challenge:-random_noise\" data-toc-modified-id=\"Code-Challenge:-random_noise-1.3\"><span class=\"toc-item-num\">1.3&nbsp;&nbsp;</span>Code Challenge: <code>random_noise</code></a></span></li><li><span><a href=\"#Code-Challenge:-mirror_image\" data-toc-modified-id=\"Code-Challenge:-mirror_image-1.4\"><span class=\"toc-item-num\">1.4&nbsp;&nbsp;</span>Code Challenge: <code>mirror_image</code></a></span></li><li><span><a href=\"#Code-Challenge:-Augmenting-Data\" data-toc-modified-id=\"Code-Challenge:-Augmenting-Data-1.5\"><span class=\"toc-item-num\">1.5&nbsp;&nbsp;</span>Code Challenge: Augmenting Data</a></span></li></ul></li><li><span><a href=\"#Extensions\" data-toc-modified-id=\"Extensions-2\"><span class=\"toc-item-num\">2&nbsp;&nbsp;</span>Extensions</a></span></li></ul></div>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Data Augmentation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Suppose we'll be training a binary classifier to distinguish between images of cats and dogs. We've got a data set of a few thousand images of furry beasts but we'd like to augment our data set in the hopes of training a more generalizable model. We'll use this time to explore some of the image processing features of `skimage`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import skimage as sk\n",
+    "from skimage import transform\n",
+    "from skimage import util\n",
+    "import numpy as np"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Code Challenge: Importing Images"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "There are zipped files containing images of cats and dogs in the directory `data/cats-and-dogs`. Unzip them and use `sk.io.imread` to try reading a few of the images. What data type is the resulting in-memory representation of the image?"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# try reading some images of cats and dogs"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Code Challenge: `random_rotation`"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Using `sk.transform.rotate`, create a function that randomly rotates an image at an angle between -30 and 30 degrees."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def random_rotate(im_array):\n",
+    "    pass"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Code Challenge: `random_noise`"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Using `sk.util.random_noise`, create a function that adds random noise to an image."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def random_noise(im_array):\n",
+    "    pass"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Code Challenge: `mirror_image`"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Write a function that flips an image around the vertical axis through the center of the image. You shouldn't need any libraries."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def mirror_image(im_array):\n",
+    "    pass"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Code Challenge: Augmenting Data"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Come up with a plan to double the size of your data set using the functions you've created. Not every image should be treated the same way (i.e. don't apply each transformation to each image). Use `sk.io.imsave` to save your transformed images. The filenames you use should follow the format for the pre-existing images."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Use this cell to perform data augmentation"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Extensions"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Explore more of the preprocessing functionality in the `skimage` library. Are there any other transformations you want to apply to your data? You may also return to your previous unfinished hack hours."
+   ]
+  }
+ ],
+ "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": "3.6.4"
+  },
+  "toc": {
+   "nav_menu": {},
+   "number_sections": true,
+   "sideBar": false,
+   "skip_h1_title": false,
+   "title_cell": "Table of Contents",
+   "title_sidebar": "Contents",
+   "toc_cell": true,
+   "toc_position": {},
+   "toc_section_display": true,
+   "toc_window_display": false
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}