deepdraw-TSN-RGB.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualizing Action Classes Learned with RGB Flow Models\n",
    "\n",
    "This notebook is adapted from the original DeepDraw notebook. In particular, this notebook generates RGB images from action recognition models trained on RGB frames. The output is a image refelecting the knowledge in the model of a specific action class.\n",
    "\n",
    "------\n",
    "\n",
    "\n",
    "This is an ipython notebook to generate visualizations of classes with GoogLeNet, for some more info refer to [this blogpost](http://auduno.com/post/125362849838/visualizing-googlenet-classes), and for some examples of generated images see [this](https://goo.gl/photos/8qcvjnYBQVSGG2eN6) album of highlights or [this](https://goo.gl/photos/FfsZZektqpZkdDnKA) album of all 1000 imagenet classes.\n",
    "\n",
    "To run this code, you'll need an installation of Caffe with built pycaffe libraries, as well as the python libraries numpy, scipy and PIL. For instructions on how to install Caffe and pycaffe, refer to the installation guide [here](http://caffe.berkeleyvision.org/installation.html). Before running the ipython notebooks, you'll also need to download the [bvlc_googlenet model](https://github.com/BVLC/caffe/tree/master/models/bvlc_googlenet), and modify the variables ```pycaffe_root``` to refer to the path of your pycaffe installation (if it's not already in your python path) and ```model_path``` to refer to the path of the googlenet caffe model. Also uncomment the line that enables GPU mode if you have built Caffe with GPU-support and a suitable GPU available.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "caffe_path = '/path/to/caffe_root'\n",
    "tsn_path = '/path/to/tsn'\n",
    "save_path = ''\n",
    "# which action class to visualize (0-based), to view the class name, see <tsn_path>/data/ucf101_splits/classInd.txt\n",
    "action_class = 48"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# imports and basic notebook setup\n",
    "from cStringIO import StringIO\n",
    "import numpy as np\n",
    "import os,re,random\n",
    "import scipy.ndimage as nd\n",
    "import PIL.Image\n",
    "import sys\n",
    "from IPython.display import clear_output, Image, display\n",
    "from scipy.misc import imresize\n",
    "\n",
    "pycaffe_root = caffe_path+\"/python\" \n",
    "sys.path.insert(0, pycaffe_root)\n",
    "import caffe\n",
    "\n",
    "model_name = \"TSN-BN-Inception-RGB\"\n",
    "net_fn   = './tsn_bn_inception_rgb_deploy.prototxt'\n",
    "param_fn = tsn_path+'models/ucf101_split1_tsn_rgb_reference_bn_inception.caffemodel'\n",
    "mean = np.float32([104.0, 117.0, 123.0])\n",
    "\n",
    "caffe.set_mode_gpu() # uncomment this if gpu processing is available\n",
    "caffe.set_device(0)\n",
    "net = caffe.Classifier(net_fn, param_fn,\n",
    "                       mean = mean, # ImageNet mean, training set dependent\n",
    "                       channel_swap = (2,1,0)) # the reference model has channels in BGR order instead of RGB\n",
    "\n",
    "# a couple of utility functions for converting to and from Caffe's input image layout\n",
    "def preprocess(net, img):\n",
    "    return np.float32(np.rollaxis(img, 2)[::-1]) - net.transformer.mean['data']\n",
    "def deprocess(net, img):\n",
    "    return np.dstack((img + net.transformer.mean['data'])[::-1])\n",
    "\n",
    "def blur(img, sigma):\n",
    "    if sigma > 0:\n",
    "        img[0] = nd.filters.gaussian_filter(img[0], sigma, order=0)\n",
    "        img[1] = nd.filters.gaussian_filter(img[1], sigma, order=0)\n",
    "        img[2] = nd.filters.gaussian_filter(img[2], sigma, order=0)\n",
    "    return img\n",
    "\n",
    "def showarray(a, f, fmt='jpeg'):\n",
    "    a = np.uint8(np.clip(a, 0, 255))\n",
    "    f = StringIO()\n",
    "    PIL.Image.fromarray(a).save(f, fmt)\n",
    "    display(Image(data=f.getvalue()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Definition of the main gradient ascent functions. Note that these are based on the [deepdream code](https://github.com/google/deepdream/blob/master/dream.ipynb) published by Google as well as [this code](https://github.com/kylemcdonald/deepdream/blob/master/dream.ipynb) by Kyle McDonald."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def make_step(net, step_size=1.5, end='inception_4c/output', clip=True, focus=None, sigma=None):\n",
    "    '''Basic gradient ascent step.'''\n",
    "\n",
    "    src = net.blobs['data'] # input image is stored in Net's 'data' blob\n",
    "    \n",
    "    dst = net.blobs[end]\n",
    "    net.forward(end=end)\n",
    "\n",
    "    one_hot = np.zeros_like(dst.data)\n",
    "    one_hot.flat[focus] = 1.\n",
    "    dst.diff[:] = one_hot\n",
    "\n",
    "    net.backward(start=end)\n",
    "    g = src.diff[0]\n",
    "    \n",
    "    src.data[:] += step_size/np.abs(g).mean() * g\n",
    "\n",
    "    if clip:\n",
    "        bias = net.transformer.mean['data']\n",
    "        src.data[:] = np.clip(src.data, -bias, 255-bias) \n",
    "        \n",
    "    src.data[0] = blur(src.data[0], sigma)\n",
    "    \n",
    "    # reset objective for next step\n",
    "    dst.diff.fill(0.)\n",
    "\n",
    "def deepdraw(net, base_img, octaves, random_crop=True, visualize=True, focus=None,\n",
    "    clip=True, **step_params):\n",
    "    \n",
    "    # prepare base image\n",
    "    image = preprocess(net, base_img) # (3,224,224)\n",
    "\n",
    "    # get input dimensions from net\n",
    "    w = net.blobs['data'].width\n",
    "    h = net.blobs['data'].height\n",
    "    \n",
    "    print \"starting drawing\"\n",
    "    src = net.blobs['data']\n",
    "    src.reshape(1,3,h,w) # resize the network's input image size\n",
    "    for e,o in enumerate(octaves):\n",
    "        if 'scale' in o:\n",
    "            # resize by o['scale'] if it exists\n",
    "            image = nd.zoom(image, (1,o['scale'],o['scale']))\n",
    "        _,imw,imh = image.shape\n",
    "        \n",
    "        # select layer\n",
    "        layer = o['layer']\n",
    "\n",
    "        for i in xrange(o['iter_n']):\n",
    "            if imw > w:\n",
    "                if random_crop:\n",
    "                    # randomly select a crop \n",
    "                    #ox = random.randint(0,imw-224)\n",
    "                    #oy = random.randint(0,imh-224)\n",
    "                    mid_x = (imw-w)/2.\n",
    "                    width_x = imw-w\n",
    "                    ox = np.random.normal(mid_x, width_x*0.3, 1)\n",
    "                    ox = int(np.clip(ox,0,imw-w))\n",
    "                    mid_y = (imh-h)/2.\n",
    "                    width_y = imh-h\n",
    "                    oy = np.random.normal(mid_y, width_y*0.3, 1)\n",
    "                    oy = int(np.clip(oy,0,imh-h))\n",
    "                    # insert the crop into src.data[0]\n",
    "                    src.data[0] = image[:,ox:ox+w,oy:oy+h]\n",
    "                else:\n",
    "                    ox = (imw-w)/2.\n",
    "                    oy = (imh-h)/2.\n",
    "                    src.data[0] = image[:,ox:ox+w,oy:oy+h]\n",
    "            else:\n",
    "                ox = 0\n",
    "                oy = 0\n",
    "                src.data[0] = image.copy()\n",
    "                \n",
    "            sigma = o['start_sigma'] + ((o['end_sigma'] - o['start_sigma']) * i) / o['iter_n']\n",
    "            step_size = o['start_step_size'] + ((o['end_step_size'] - o['start_step_size']) * i) / o['iter_n']\n",
    "            \n",
    "            make_step(net, end=layer, clip=clip, focus=focus, \n",
    "                      sigma=sigma, step_size=step_size)\n",
    "\n",
    "            if visualize:\n",
    "                vis = deprocess(net, src.data[0])\n",
    "                if not clip: # adjust image contrast if clipping is disabled\n",
    "                    vis = vis*(255.0/np.percentile(vis, 99.98))\n",
    "                if i % 1 == 0:\n",
    "                    showarray(vis,\"./filename\"+str(i)+\".jpg\")\n",
    "            \n",
    "            if i % 50 == 0 or i ==o['iter_n']-1:\n",
    "                print 'finished step %d in octave %d' % (i,e)\n",
    "            \n",
    "            # insert modified image back into original image (if necessary)\n",
    "            image[:,ox:ox+w,oy:oy+h] = src.data[0]\n",
    "        \n",
    "        print \"octave %d image:\" % e\n",
    "        \n",
    "        showarray(deprocess(net, image),\"./octave_\"+str(e)+\".jpg\")\n",
    "            \n",
    "    # returning the resulting image\n",
    "    return deprocess(net, image)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Generating the class visualizations\n",
    "\n",
    "The ```octaves``` list determines in which order we optimize layers, as well as how many iterations and scaling on each octave. For each octave, parameters are:\n",
    "* ```layer``` : which layer to optimize\n",
    "* ```iter_n``` : how many iterations\n",
    "* ```scale``` : by what factor (if any) to scale up the base image before proceeding\n",
    "* ```start_sigma``` : the initial radius of the gaussian blur\n",
    "* ```end_sigma``` : the final radius of the gaussian blur\n",
    "* ```start_step_size``` : the initial step size of the gradient ascent\n",
    "* ```end_step_size``` : the final step size of the gradient ascent\n",
    "\n",
    "The choice of octave parameters below will give decent images, and is the one used for visualizations in the blogpost. However, the choice of parameters was a bit arbitrary, so feel free to experiment. Note that generating an image will take around 1 minute with GPU-enabled Caffe, or 10-15 minutes if you're running purely on CPU, depending on your computer performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "starting drawing\n",
      "finished step 0 in octave 0\n",
      "finished step 50 in octave 0\n",
      "finished step 100 in octave 0\n",
      "finished step 150 in octave 0\n",
      "finished step 189 in octave 0\n",
      "octave 0 image:\n"
     ]
    },
    {
     "data": {
      "image/jpeg": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0a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\n9kEcShTk56nvUq26pxzxyaV+pzOTZ//Z\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "finished step 0 in octave 1\n",
      "finished step 50 in octave 1\n",
      "finished step 100 in octave 1\n",
      "finished step 149 in octave 1\n",
      "octave 1 image:\n"
     ]
    },
    {
     "data": {
      "image/jpeg": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0a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\n1NBm7s//2Q==\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "finished step 0 in octave 2\n",
      "finished step 50 in octave 2\n",
      "finished step 100 in octave 2\n",
      "finished step 149 in octave 2\n",
      "octave 2 image:\n"
     ]
    },
    {
     "data": {
      "image/jpeg": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0a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\n0SqAGwOlWY/vLRRSexDHH/W0mBgjHU0UVmApABoooqgP/9k=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "finished step 0 in octave 3\n",
      "finished step 9 in octave 3\n",
      "octave 3 image:\n"
     ]
    },
    {
     "data": {
      "image/jpeg": "/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0a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\nx/1tJgYIx1NFFZgKQAaKKKoD/9k=\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# these octaves determine gradient ascent steps\n",
    "octaves = [\n",
    "    {\n",
    "        'layer':'fc-action',\n",
    "        'iter_n':190,\n",
    "        'start_sigma':2.5,\n",
    "        'end_sigma':0.78,\n",
    "        'start_step_size':11.,\n",
    "        'end_step_size':11.\n",
    "    },\n",
    "    {\n",
    "        'layer':'fc-action',\n",
    "        'scale':1.2,\n",
    "        'iter_n':150,\n",
    "        'start_sigma':0.78*1.2,\n",
    "        'end_sigma':0.5,\n",
    "        'start_step_size':6.,\n",
    "        'end_step_size':6.\n",
    "    },\n",
    "    {\n",
    "        'layer':'fc-action',\n",
    "        'scale':1.2,\n",
    "        'iter_n':150,\n",
    "        'start_sigma':0.78*1.2,\n",
    "        'end_sigma':0.44,\n",
    "        'start_step_size':6.,\n",
    "        'end_step_size':3.\n",
    "    },\n",
    "    {\n",
    "        'layer':'fc-action',\n",
    "        'iter_n':10,\n",
    "        'start_sigma':0.44,\n",
    "        'end_sigma':0.304,\n",
    "        'start_step_size':3.,\n",
    "        'end_step_size':3.\n",
    "    }\n",
    "]\n",
    "\n",
    "# get original input size of network\n",
    "original_w = net.blobs['data'].width\n",
    "original_h = net.blobs['data'].height\n",
    "# the background color of the initial image\n",
    "background_color = np.float32([200.0, 200.0, 200.0])\n",
    "# generate initial random image\n",
    "gen_image = np.random.normal(background_color, 12, (original_w, original_h, 3))\n",
    "\n",
    "\n",
    "\n",
    "# generate class visualization via octavewise gradient ascent\n",
    "gen_image = deepdraw(net, gen_image, octaves, focus=action_class, \n",
    "                 random_crop=True, visualize=False)\n",
    "\n",
    "# save image\n",
    "import os\n",
    "img_fn = os.path.join(save_path, '_'.join([\"tsn_visualization_rgb\", str(action_class)+'.png']))\n",
    "PIL.Image.fromarray(np.uint8(gen_image)).save(img_fn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}