{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Project 2\n",
    "## Image Classification with CNN\n",
    "-------------------------------\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Direct prediction\n",
    "   * Download the dataset birds from http://www-cvr.ai.uiuc.edu/ponce_grp/data/.\n",
    "   * Use [Keras](https://keras.io/) and a CNN from Keras [Applications](https://keras.io/applications/) pretrained on ImageNet, to classify the images in the birds dataset. Construct a confusion matrix that relates the bird classes with the 10 most frequent classes from ImageNet predicted by the model.\n",
    "   * Discuss the results.\n",
    "   \n",
    "## 2. Transfer learning\n",
    "   * Use the pre-trained CNN model as a feature extractor. Create a new model that replaces the top part of the pretrained CNN with two layers of 256 and 6 neurons respectively.\n",
    "   * Train the model with the training images from the bird dataset. \n",
    "   * Evaluate the performance over the test dataset reporting the results in a confusion matrix. Discuss the results. \n",
    "   \n",
    "## 3. Fine tuning\n",
    "   * Repeat the experiment from the last question, but this time allow all the layers to be trained. \n",
    "   * Compare and discuss the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}