{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "import os\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "import glob\n",
    "import cPickle as pickle\n",
    "import PIL\n",
    "from PIL import Image\n",
    "import random\n",
    "\n",
    "import caffe\n",
    "\n",
    "from lib import run_net\n",
    "from lib import score_util\n",
    "from lib import plot_util\n",
    "\n",
    "from datasets.youtube import youtube\n",
    "from datasets.pascal_voc import pascal\n",
    "\n",
    "plt.rcParams['image.cmap'] = 'gray'\n",
    "plt.rcParams['image.interpolation'] = 'nearest'\n",
    "plt.rcParams['figure.figsize'] = (12, 12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Configure Caffe. **Only necessary for collecting differences, otherwise skip.** "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "caffe.set_device(0)\n",
    "caffe.set_mode_gpu()\n",
    "\n",
    "net = caffe.Net('../nets/voc-fcn8s.prototxt', '../nets/voc-fcn8s-heavy.caffemodel', caffe.TEST)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dataset details"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "PV = pascal('/x/PASCAL/VOC2011')\n",
    "YT = youtube('/x/youtube/')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Frame Differencing\n",
    "\n",
    "Extract fcn8s on all of the frames that belong to labeled youtube videos for analysis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def keep_layer(layer):\n",
    "    \"\"\"\n",
    "    Filter layers to keep scores alone\n",
    "    \"\"\"\n",
    "    is_split = 'split' in layer\n",
    "    is_crop = 'c' == layer[-1]\n",
    "    is_interp = 'up' in layer\n",
    "    is_type = layer.startswith('score')\n",
    "    return is_type and not (is_split or is_crop or is_interp)\n",
    "\n",
    "def load_layer_diffs(class_, vid, shot):\n",
    "    diffs = np.load('{}/v1/{}/data/{}/shots/{}/diffs.npz'.format(YT.dir, class_, vid, shot))\n",
    "    layers = diffs.keys()\n",
    "    diffs = np.concatenate([d[..., np.newaxis] for l, d in diffs.iteritems()], axis=-1)\n",
    "    return layers, diffs\n",
    "    \n",
    "def segsave(net, path):\n",
    "    \"\"\"\n",
    "    Save maps to disk as compressed arrays.\n",
    "    \"\"\"\n",
    "    for layer, blob in net.blobs.iteritems():\n",
    "        if keep_layer(layer):\n",
    "            np.savez('{}-{}.npz'.format(path, layer), blob.data[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run through the frames, difference, and store\n",
    "\n",
    "- differences in scores (for confidence)\n",
    "- differences in actual output (proportion of pixels changed)\n",
    "- differences in data, differences in label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "layers to difference: ['score_fr', 'score_pool4', 'score_pool3', 'score']\n",
      "differencing aeroplane\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0002\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0013\n",
      "\tskipping...\n",
      "differencing bird\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0007\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "differencing boat\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0004\n",
      "\tskipping...\n",
      "\tvid 0005\n",
      "\tskipping...\n",
      "\tvid 0006\n",
      "\tskipping...\n",
      "\tvid 0007\n",
      "\tskipping...\n",
      "\tvid 0008\n",
      "\tskipping...\n",
      "\tvid 0009\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "\tvid 0015\n",
      "\tskipping...\n",
      "\tvid 0016\n",
      "\tskipping...\n",
      "\tvid 0017\n",
      "\tskipping...\n",
      "differencing car\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0002\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0004\n",
      "\tskipping...\n",
      "\tvid 0005\n",
      "\tskipping...\n",
      "\tvid 0008\n",
      "\tskipping...\n",
      "\tvid 0009\n",
      "\tskipping...\n",
      "differencing cat\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0002\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0004\n",
      "\tskipping...\n",
      "\tvid 0006\n",
      "\tskipping...\n",
      "\tvid 0008\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0013\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "\tvid 0015\n",
      "\tskipping...\n",
      "\tvid 0016\n",
      "\tskipping...\n",
      "\tvid 0017\n",
      "\tskipping...\n",
      "\tvid 0018\n",
      "\tskipping...\n",
      "\tvid 0020\n",
      "\tskipping...\n",
      "differencing cow\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0002\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0004\n",
      "\tskipping...\n",
      "\tvid 0005\n",
      "\tskipping...\n",
      "\tvid 0006\n",
      "\tskipping...\n",
      "\tvid 0007\n",
      "\tskipping...\n",
      "\tvid 0008\n",
      "\tskipping...\n",
      "\tvid 0009\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0013\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "\tvid 0015\n",
      "\tskipping...\n",
      "\tvid 0016\n",
      "\tskipping...\n",
      "\tvid 0017\n",
      "\tskipping...\n",
      "\tvid 0018\n",
      "\tskipping...\n",
      "\tvid 0019\n",
      "\tskipping...\n",
      "\tvid 0022\n",
      "\tskipping...\n",
      "differencing dog\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0005\n",
      "\tskipping...\n",
      "\tvid 0006\n",
      "\tskipping...\n",
      "\tvid 0007\n",
      "\tskipping...\n",
      "\tvid 0008\n",
      "\tskipping...\n",
      "\tvid 0009\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0013\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "\tvid 0016\n",
      "\tskipping...\n",
      "\tvid 0018\n",
      "\tskipping...\n",
      "\tvid 0020\n",
      "\tskipping...\n",
      "\tvid 0021\n",
      "\tskipping...\n",
      "\tvid 0022\n",
      "\tskipping...\n",
      "\tvid 0023\n",
      "\tskipping...\n",
      "\tvid 0025\n",
      "\tskipping...\n",
      "\tvid 0026\n",
      "\tskipping...\n",
      "\tvid 0027\n",
      "\tskipping...\n",
      "\tvid 0028\n",
      "\tskipping...\n",
      "\tvid 0030\n",
      "\tskipping...\n",
      "\tvid 0031\n",
      "\tskipping...\n",
      "\tvid 0032\n",
      "\tskipping...\n",
      "\tvid 0034\n",
      "\tskipping...\n",
      "\tvid 0035\n",
      "\tskipping...\n",
      "\tvid 0036\n",
      "\tskipping...\n",
      "differencing horse\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0009\n",
      "\tskipping...\n",
      "\tvid 0010\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "\tvid 0018\n",
      "\tskipping...\n",
      "\tvid 0020\n",
      "\tskipping...\n",
      "\tvid 0021\n",
      "\tskipping...\n",
      "\tvid 0022\n",
      "\tskipping...\n",
      "\tvid 0024\n",
      "\tskipping...\n",
      "\tvid 0025\n",
      "\tskipping...\n",
      "\tvid 0026\n",
      "\tskipping...\n",
      "\tvid 0029\n",
      "\tskipping...\n",
      "differencing motorbike\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0002\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0006\n",
      "\tskipping...\n",
      "\tvid 0007\n",
      "\tskipping...\n",
      "\tvid 0009\n",
      "\tskipping...\n",
      "\tvid 0011\n",
      "\tskipping...\n",
      "\tvid 0012\n",
      "\tskipping...\n",
      "\tvid 0013\n",
      "\tskipping...\n",
      "\tvid 0014\n",
      "\tskipping...\n",
      "differencing train\n",
      "\tvid 0001\n",
      "\tskipping...\n",
      "\tvid 0003\n",
      "\tskipping...\n",
      "\tvid 0008\n",
      "\tskipping...\n",
      "\tvid 0024\n",
      "\tskipping...\n",
      "\tvid 0025\n",
      "\tskipping...\n"
     ]
    }
   ],
   "source": [
    "# decide layers to difference\n",
    "layers = filter(lambda l: keep_layer(l), net.blobs.keys())\n",
    "print 'layers to difference: {}'.format(layers)\n",
    "\n",
    "# collect frame differences\n",
    "for class_ in YT.classes[1:]:  # skip background\n",
    "    print 'differencing {}'.format(class_)\n",
    "    for vid in sorted(YT.list_label_vids(class_)):\n",
    "        print '\\tvid {}'.format(vid)\n",
    "        sys.stdout.flush()\n",
    "        for shot in YT.list_label_shots(class_, vid):\n",
    "            # skip if already done\n",
    "            if os.path.exists('{}/v1/{}/data/{}/shots/{}/diffs.npz'.format(YT.dir, class_, vid, shot)):\n",
    "                print '\\tskipping...'\n",
    "                continue\n",
    "            frames = YT.list_frames(class_, vid, shot)\n",
    "            label_frames = YT.list_label_frames(class_, vid, shot)\n",
    "            # seek to beginning of labels\n",
    "            first_label_idx = int(label_frames[0]) - 1\n",
    "            frames = frames[first_label_idx:]\n",
    "            # handle first frame\n",
    "            data = YT.preprocess(YT.load_frame(class_, vid, shot, frames[0]))\n",
    "            label = YT.convert_yt2voc_label(YT.load_label(class_, vid, shot, frames[0]), class_, PV.classes)\n",
    "            run_net.segrun(net, data)\n",
    "            feats = [net.blobs[l].data[0].copy() for l in layers]\n",
    "            argmaxes = [net.blobs[l].data[0].argmax(axis=0).copy() for l in layers if 'score' in l]\n",
    "            # differences: layers, then argmaxes, and last is data and label\n",
    "            diffs = {}\n",
    "            zeros = np.zeros((len(frames)), dtype=np.float32)\n",
    "            diffs['data'] = zeros.copy()\n",
    "            diffs['label'] = zeros.copy()\n",
    "            for l in layers:\n",
    "                diffs[l] = zeros.copy()\n",
    "                diffs[l + '-argmax'] = zeros.copy()\n",
    "            for ix, frame in enumerate(frames):\n",
    "                # extract this frame\n",
    "                new_data = YT.preprocess(YT.load_frame(class_, vid, shot, frame))\n",
    "                if frame in label_frames:\n",
    "                    new_label = YT.convert_yt2voc_label(YT.load_label(class_, vid, shot, frame), class_, PV.classes)\n",
    "                run_net.segrun(net, new_data)\n",
    "                new_feats = [net.blobs[l].data[0].copy() for l in layers]\n",
    "                new_argmaxes = [net.blobs[l].data[0].argmax(axis=0).copy() for l in layers if 'score' in l]\n",
    "                # compute differences\n",
    "                for lx, l in enumerate(layers):\n",
    "                    abs_diff = np.abs(new_feats[lx] - feats[lx])\n",
    "                    abs_avg = (np.abs(new_feats[lx]) + np.abs(feats[lx])) / 2.\n",
    "                    rel_diff = abs_diff / abs_avg\n",
    "                    rel_diff[np.isnan(rel_diff)] = 0\n",
    "                    diffs[l][ix] = rel_diff.mean()\n",
    "                    diffs[l + '-argmax'][ix] = np.array(new_argmaxes[lx] != argmaxes[lx]).mean()\n",
    "                diffs['data'][ix] = (np.abs(new_data - data) / ((np.abs(new_data) + np.abs(data)) / 2.)).mean()\n",
    "                diffs['label'][ix] = np.array(new_label != label).mean()\n",
    "                # advance over old\n",
    "                feats, argmaxes, data, label = new_feats, new_argmaxes, new_data, new_label\n",
    "            np.savez('{}/v1/{}/data/{}/shots/{}/diffs.npz'.format(YT.dir, class_, vid, shot), **diffs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inspection\n",
    "\n",
    "Once the differences have been computed offline and saved, we can inspect the traces.\n",
    "Start by checking the mean across layers of a given shot."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score-argmax        : 0.0235913135111\n",
      "score_pool3-argmax  : 0.173584058881\n",
      "score_fr            : 0.282327950001\n",
      "label               : 0.0078412136063\n",
      "score_pool4         : 0.316808968782\n",
      "score               : 0.290756285191\n",
      "score_pool3         : 0.2999342978\n",
      "score_fr-argmax     : 0.0252976194024\n",
      "data                : 0.351404070854\n",
      "score_pool4-argmax  : 0.140458166599\n"
     ]
    }
   ],
   "source": [
    "# show mean differences across layers\n",
    "layers, diffs = load_layer_diffs('boat', '0011', '001')\n",
    "for layer, diff in zip(layers, diffs.T):\n",
    "    print '{:<20}: {}'.format(layer, np.mean(diff))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Grand mean and standard deviation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score-argmax        : 0.0206552718364\n",
      "score_pool3-argmax  : 0.0546926715348\n",
      "score_fr            : 0.120616351528\n",
      "label               : 0.00458718119848\n",
      "score_pool4         : 0.102406660096\n",
      "score               : 0.12047903273\n",
      "score_pool3         : 0.093659251823\n",
      "score_fr-argmax     : 0.0206845826048\n",
      "data                : 0.178912213557\n",
      "score_pool4-argmax  : 0.0592727472488\n"
     ]
    }
   ],
   "source": [
    "all_diffs = []\n",
    "for class_ in YT.classes:\n",
    "    for vid in sorted(YT.list_label_vids(class_)):\n",
    "        for shot in YT.list_label_shots(class_, vid):\n",
    "            layers, diffs = load_layer_diffs(class_, vid, shot)\n",
    "            all_diffs.append(diffs)\n",
    "all_diff_arr = np.concatenate(all_diffs)\n",
    "\n",
    "means = np.zeros((len(all_diffs), len(layers)))\n",
    "for ix, diff in enumerate(all_diffs):\n",
    "    means[ix] = np.mean(diff, axis=0)\n",
    "\n",
    "for layer, mean in zip(layers, means.T):\n",
    "    print '{:<20}: {}'.format(layer, np.std(mean))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot difference traces for a random video"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dog 0009 001\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f8c342341d0>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtQAAALICAYAAABfOWcQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nagAAGCCoAQBgwP8ByiDkCxkYYXcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8c34234e90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_ = random.choice(YT.classes)\n",
    "vid = random.choice(YT.list_label_vids(class_))\n",
    "shot = random.choice(YT.list_label_shots(class_, vid))\n",
    "print class_, vid, shot\n",
    "\n",
    "layers, diffs = load_layer_diffs(class_, vid, shot)\n",
    "diff_means = np.mean(diffs, axis=0)\n",
    "\n",
    "plot_layers = [l for l in layers if 'argmax' in l] + ['label']\n",
    "plot_ix = [layers.index(l) for l in plot_layers]\n",
    "\n",
    "plt.figure()\n",
    "plt.title('{} vid {} shot {}'.format(class_, vid, shot))\n",
    "plt.plot(diffs[:, plot_ix] - diff_means[plot_ix])\n",
    "plt.legend(plot_layers)"
   ]
  }
 ],
 "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": "2.7.11+"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}