{
 "metadata": {
  "name": "02 Compute Gabor Filter Responses"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy\n",
      "import math\n",
      "import scipy\n",
      "import scipy.io\n",
      "import scipy.misc\n",
      "import scipy.cluster\n",
      "import scipy.cluster.vq\n",
      "import matplotlib\n",
      "import matplotlib.pyplot\n",
      "import os\n",
      "import IPython\n",
      "import IPython.parallel\n",
      "import itertools\n",
      "import random\n",
      "import sklearn\n",
      "import sklearn.decomposition\n",
      "import sklearn.manifold\n",
      "import sklearn.cluster\n",
      "import sklearn.feature_extraction\n",
      "import sklearn.feature_extraction.text\n",
      "\n",
      "base_path = \"/u/mlrobert/code/local/2013_drawing_assistant/data/\""
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "num_thetas  = 8\n",
      "filter_path = base_path+\"filters/gabor/num_thetas=08/\"\n",
      "\n",
      "tmp = []\n",
      "\n",
      "for theta_index in range(num_thetas):\n",
      "    \n",
      "    current_g_mat = scipy.io.loadmat(filter_path+\"theta=%02d.mat\" % theta_index)\n",
      "    current_g     = current_g_mat[\"g\"]\n",
      "    \n",
      "    tmp.append(current_g)\n",
      "\n",
      "g = numpy.array(tmp)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "img_path            = base_path+\"renderings/bunny2/apparent_ridges.num_latitude_lines=20.num_longitude_lines=20/\"\n",
      "responses_base_path = base_path+\"filter_responses/bunny2/apparent_ridges.num_latitude_lines=20.num_longitude_lines=20.gabor.num_thetas=08/\"\n",
      "debug_base_path     = base_path+\"debug/filter_responses/bunny2/apparent_ridges.num_latitude_lines=20.num_longitude_lines=20.gabor.num_thetas=08/\""
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "def compute_gabor_filter_responses(img_name_ext):\n",
      "    \n",
      "    import numpy\n",
      "    import scipy\n",
      "    import scipy.io\n",
      "    import scipy.misc\n",
      "    import os\n",
      "    import matplotlib\n",
      "    import matplotlib.pyplot\n",
      "\n",
      "    img_name = img_name_ext[:-4]\n",
      "    img_ext  = img_name_ext[-4:]\n",
      "    \n",
      "    img = scipy.misc.imread(img_path+img_name+img_ext).astype(numpy.float32)[:,:,0]\n",
      "    img = (255.0 - img) / 255.0\n",
      "    \n",
      "    responses_path = responses_base_path+img_name+\"/\"\n",
      "    debug_path = debug_base_path+img_name+\"/\"\n",
      "    \n",
      "    if not os.path.exists(responses_path):\n",
      "        os.makedirs(responses_path)\n",
      "\n",
      "    if not os.path.exists(debug_path):\n",
      "        os.makedirs(debug_path)\n",
      "        \n",
      "    for theta_index in range(num_thetas):\n",
      "        \n",
      "        F_img = numpy.fft.fft2(img)\n",
      "        F_img_centered = numpy.fft.fftshift(F_img)\n",
      "        F_img_centered_filtered = F_img_centered * g[theta_index]\n",
      "        F_img_centered_filtered_uncentered = numpy.fft.ifftshift(F_img_centered_filtered)\n",
      "        filter_response = numpy.fft.ifft2(F_img_centered_filtered_uncentered)\n",
      "        \n",
      "        scipy.io.savemat(\"%stheta=%02d.mat\" % (responses_path, theta_index), {\"filter_response\":filter_response}, oned_as=\"column\")\n",
      "        \n",
      "        matplotlib.pyplot.clf();\n",
      "        matplotlib.pyplot.imshow(numpy.abs(filter_response), interpolation=\"nearest\");\n",
      "        matplotlib.pyplot.colorbar();\n",
      "        matplotlib.pyplot.savefig(\"%stheta=%02d.png\" % (debug_path, theta_index));\n",
      "        matplotlib.pyplot.clf();"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 46
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "img_names_exts = sorted(os.listdir(img_path))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 47
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "compute_gabor_filter_responses(img_names_exts[0])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 48
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "client = IPython.parallel.Client()\n",
      "directView = client[:]\n",
      "\n",
      "directView[\"num_thetas\"]          = num_thetas\n",
      "directView[\"img_path\"]            = img_path\n",
      "directView[\"responses_base_path\"] = responses_base_path\n",
      "directView[\"debug_path\"]          = debug_path\n",
      "directView[\"g\"]                   = g\n",
      "\n",
      "asyncResult = directView.map_async(compute_gabor_filter_responses, img_names_exts)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 49
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print asyncResult.progress\n",
      "print asyncResult.ready()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "32\n",
        "True\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "result = asyncResult.get()\n",
      "print result"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None]\n"
       ]
      }
     ],
     "prompt_number": 35
    }
   ],
   "metadata": {}
  }
 ]
}