Example usage.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### An example notebook on how to use the Eyetracking repo"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is copied from a recent driving session. In the driving sessions, each run includes an eyetracking calibration sequence at the beginning, and this notebook is written around this experiment structure.\n",
    "\n",
    "For cases in which the calibration video is separate from the data video, see step 3.\n",
    "\n",
    "Important classes in this package:\n",
    "* `PupilFinder` (and child classes): handles finding the pupil in videos. Child classes implement different algorithms.\n",
    "* `EyetrackingClibrator`: takes pupil traces and estimates a mapping from pupil location in video space to gaze location in screen space. It includes a `PupilFinder` field that does this.\n",
    "* `Eyetracker`: an unfinished top level class that combines things and gets you things that you can get features from.\n",
    "\n",
    "Full documentation can be found in the [Docs folder](./Docs/index.html) of this repo."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-18T00:37:42.155422Z",
     "start_time": "2020-08-18T00:37:39.744778Z"
    }
   },
   "outputs": [],
   "source": [
    "from Eyetracking import EyetrackingCalibrator, Eyetracker, TemplatePupilFinder, EyetrackingUtilities\n",
    "import numpy\n",
    "from skimage.transform import warp, AffineTransform\n",
    "from skimage import io\n",
    "import cv2\n",
    "from glob import glob"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Take a quick look at the video. We use this to define a sub-window in the video to search for the pupil, since runtime is directly proportional to the size of the search area."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T19:51:30.241320Z",
     "start_time": "2020-08-17T19:51:30.001261Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "<matplotlib.image.AxesImage at 0x7f02e8e89b90>"
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": "<matplotlib.figure.Figure at 0x7f02e8eed390>"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "video = cv2.VideoCapture('/E/Driving Data/20200814TZ/Eyetracking/run03.avi')\n",
    "for i in range(300):\n",
    "    r, f = video.read()\n",
    "io.imshow(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:30:25.572110Z",
     "start_time": "2020-08-17T23:30:25.566394Z"
    }
   },
   "outputs": [],
   "source": [
    "# left, right, top, bottom\n",
    "window = (25, 175, 80, 160)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To align the video to the TTLs, we make use of the history file saved by Avotec to find when the TTLs came in. Note that the time is burned into the eyetracking videos.\n",
    "Note you can also use the events file instead of the history file. To specify the type of file you are using, use the `fileType` argument to `ParseRecordsForStartTTLs`.\n",
    "The default is to use the history file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-18T00:37:43.241872Z",
     "start_time": "2020-08-18T00:37:43.226032Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "20 duplicated TTLs\n((1, 17, 36, 381), 340)\n((1, 34, 31, 689), 340)\n((1, 50, 27, 354), 340)\n((2, 10, 39, 394), 340)\n((2, 26, 27, 251), 340)\n(None, 0)\n((2, 42, 17, 911), 340)\n"
    }
   ],
   "source": [
    "starts = EyetrackingUtilities.ParseRecordsForStartTTLs('/E/Driving Data/20200814TZ/Eyetracking/20200814TZ_history.txt')\n",
    "for start in starts:\n",
    "    print(start)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`ParseHistoryForStartTTLs` finds the first TTL in each run. It does so by looking for TTLs that are separated by a large amount of time. Because sometimes Avotec records spurious signals, the found set of TTLs may be a superset of the actual TTLs and may need to be pruned by hand, as in this case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-18T00:37:43.418789Z",
     "start_time": "2020-08-18T00:37:43.411418Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "((1, 17, 36, 381), 340)\n((1, 34, 31, 689), 340)\n((1, 50, 27, 354), 340)\n((2, 10, 39, 394), 340)\n((2, 26, 27, 251), 340)\n((2, 42, 17, 911), 340)\n"
    }
   ],
   "source": [
    "starts = [starts[i] for i in [0, 1, 2 ,3 ,4, 6]]\n",
    "for start in starts:\n",
    "    print(start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:09:22.814447Z",
     "start_time": "2020-08-17T23:09:22.798736Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "['/E/Driving Data/20200814TZ/Eyetracking/run01.avi',\n '/E/Driving Data/20200814TZ/Eyetracking/run02.avi',\n '/E/Driving Data/20200814TZ/Eyetracking/run03.avi',\n '/E/Driving Data/20200814TZ/Eyetracking/run04.avi',\n '/E/Driving Data/20200814TZ/Eyetracking/run05.avi',\n '/E/Driving Data/20200814TZ/Eyetracking/run06.avi']"
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "eyetrackingFiles = glob('/E/Driving Data/20200814TZ/Eyetracking/run0*.avi')\n",
    "eyetrackingFiles.sort()\n",
    "eyetrackingFiles"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 1: finding the pupils in the videos\n",
    "\n",
    "Because I'm lazy, here's a function to deal with a video by run. The important class here is the `EyetrackingCalibrator.EyetrackingCalibrator` class. The arguments to this outer function, which are passed to the `FindPupils` method of the `EyetrackingCalibrator` class, are the 4 major parameters to  pupil finding process that will determine your success.\n",
    "\n",
    "Important arguments to constructing an `EyetrackingCalibrator` in order in which they are used here:\n",
    "* `(calibrationVideoFile): str?`: path to the eyetracking video file. This is nullable. You can pass it None and then use the Load method to load a saved calibration file, so you can skip the pupil finding process\n",
    "\n",
    "* `(calibrationBeginTime): tuple<int, int, int, int>`: Timestamp in hours, minutes, seconds, milliseconds of the first TTL in the calibration video.\n",
    "\n",
    "* `calibrationDelay: float`: number of seconds between the first TTL and the presentation\n",
    "of the first calibration dot\n",
    "\n",
    "* `calibrationDuration: float`: number of seconds that the subject fixates on\n",
    "each calibration location\n",
    "\n",
    "* `calibrationPositions: [n x 2] array`: pixel position of calibration locations on the screen. The GeneratePoints method will get you positions used in the standard eyetracking sequence.\n",
    "\n",
    "* `calibrationOrder: List[int]`: the order in which the positions in calibrationPositions are presented to the subject. In the standard eyetracking sequence, each point is presented twice, and you should use EyetrackingCalibrator.CalibrationOrder70. In the driving experiment, I present each point only once because the calibration process is repeated in every run.\n",
    "\n",
    "* `templates: bool`: Use the template matching method for finding the pupil? If false, it will use the [Hough transform](https://en.wikipedia.org/wiki/Hough_transform) to find an ellipse. That was the method we had historically used (by Shinji and Anwar, and also I think it's how Avotec finds the pupil location also), but I found that it is 1) much slower and 2) much noisier temporally than just convolving the video frame with a black circle on a white background (i.e. a pupil), and finding the position of max correlation. In the template matching method, the pupil size is found by using template circles of different sizes and picking the one with max correlation. This argument picks which `PupilFinder` class to spawn (different classes implement the two methods).\n",
    "\n",
    "Note that this function I have here writes out a video with the pupil highlighted. You should also do this when processing eyetracking videos. Check the video to make sure that the pupil is correctly circled, since that will affect the gaze estimation process. The circle should not be jumping off the pupil. In some cases, it might be pushed off to one side and squeezed by the glint (because of the correlation maths) - that is fine, because that is a stationary bias that can be taken care of by the estimator. The circle jumping around can't be.\n",
    "\n",
    "At the end of this step, the calibrator has only found where the pupil is. It has yet to estimate a gaze location.\n",
    "\n",
    "Side note: if you look in the docstrings, I use `?` to denote a nullable argument. This was written before the Python3 switch and python2 does not have type annotations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:30:26.237265Z",
     "start_time": "2020-08-17T23:30:26.224383Z"
    }
   },
   "outputs": [],
   "source": [
    "def FindPupils(run, blur = 7, minRadius = 15, maxRadius = 23, windowSize = 29):\n",
    "    \"\"\"\n",
    "    Parses one run\n",
    "    @param run:         int, run in the list of eyetracking files\n",
    "    @param blur:        int, width of the median filter used to smooth the video before finding pupils\n",
    "    @param minRadius:   int, lower bound on pupil radius to look for\n",
    "    @param maxRadius:   int, upper bound on pupil radius to look for\n",
    "    @param windowSize:  int, number of frames used to temporally smooth the pupil position\n",
    "    \"\"\"\n",
    "    # Make the calibrator object. These arguments here are most likely the ones you will need to tweak\n",
    "    calibrator = EyetrackingCalibrator.EyetrackingCalibrator(\n",
    "        eyetrackingFiles[run], \n",
    "        starts[run][0], \n",
    "        calibrationDelay = 2.0, \n",
    "        calibrationDuration = 2.0,\n",
    "        calibrationPositions = EyetrackingCalibrator.EyetrackingCalibrator.GeneratePoints(),\n",
    "        calibrationOrder = EyetrackingCalibrator.EyetrackingCalibrator.CalibrationOrder35,\n",
    "\t\tmethod = EyetrackingCalibrator.PupilFindingMethod.Templates)\n",
    "    \n",
    "    # Search the pupil in the video\n",
    "    calibrator.FindPupils(window, blur = blur, minRadius = minRadius, maxRadius = maxRadius, windowSize = windowSize)\n",
    "    \n",
    "    # Write out a copy of the eyetracking video, with a circle drawn over the pupil\n",
    "    calibrator.pupilFinder.WritePupilVideo('/E/Driving Data/20200814TZ/Eyetracking/run {} pupils.avi'.format(run + 1))\n",
    "    return calibrator"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "#### Run 1"
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:39:43.931797Z",
     "start_time": "2020-08-17T23:30:28.361134Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "240x320 (HxW) video at 60.000240001 fps with 42389 frames\n"
    }
   ],
   "source": "calibrator = FindPupils(0)"
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "I checked the video and the gaze finding looks reasonable. "
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "## Step 2: Fit a mapping function that goes from pupil location to gaze location.\nSo here we actually estimate a mapping from the pupil positions in eyetracking video space to gaze location in screen space. I will spare you the details of the process. It's a radial basis function and think-plate spline interpolater, validated using a leave-one-out cross-validation process. You can look in the `EstimateClibrationPointPositions` method for details if you so wish.\n\nLook at the `bestError` field of the `EyetrackingCalibrator` object after fitting. The is the leave-one-out error for each calibration point in which a mapping is trained on all the other points, and the pupil for the point is used to estimate gaze for that point. This is the number of pixels that the estimate is off by. Hand-wavy ranges for what this number should be:\n* 0-20: fantastic\n* 20-40: good\n* 40-60: acceptable\n* 60+: You need to tweak either the pupil finding parameters or information about the run (e.g. calibration delay, duration, time).\n\nArguments to the `Fit` method:\n* `smoothnesses: List[float]`: RBF smoothnesses - you don't need to tweak this from the defaults.\n* `methods: List[str]`: Interpolation methods - you don't need to tweak this from the defaults.\n* `varianceThreshold: float?`: If provided, the calibrator looks at how steady the fixation is at each calibration point. Points above this threshold are not used. Empirically, leave this to None (default) if the videos are good.\n* `glintVector: bool`: If true, uses the pupil-glint vector as the data, instead of just the pupil position. I empirically find this to be not helpful.\n* `searchThreshold: float`: If the calibrator finds that `bestError` is above this value, it will search a larger range of `calibrationDelay` and `calibrationDuration` values, and set itself to the combination that returns the lowest error. Setting it to 0 diables this."
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:39:51.768683Z",
     "start_time": "2020-08-17T23:39:43.970661Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "29.367180650467834"
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": "calibrator.EstimateCalibrationPointPositions()\ncalibrator.Fit()\ncalibrator.bestError"
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "This saves out pupil locations, so you just load it back in later with `calibrator.Load('path-to-file')`. Because finding pupils takes a while."
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:39:51.873994Z",
     "start_time": "2020-08-17T23:39:51.809276Z"
    }
   },
   "outputs": [],
   "source": "calibrator.Save('/E/Driving Data/20200814TZ/Eyetracking/calibration 1.dat')"
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "## Run 2"
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:49:02.315713Z",
     "start_time": "2020-08-17T23:39:51.910871Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "240x320 (HxW) video at 60.000240001 fps with 42326 frames\n"
    }
   ],
   "source": "calibrator = FindPupils(1)"
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:53:49.126495Z",
     "start_time": "2020-08-17T23:49:02.358354Z"
    }
   },
   "outputs": [],
   "source": "calibrator.FindPupils(window, blur = 11, minRadius = 15, maxRadius = 23, windowSize = 29)\ncalibrator.pupilFinder.WritePupilVideo('/E/Driving Data/20200814TZ/Eyetracking/run 2 pupils.avi')"
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:53:56.672175Z",
     "start_time": "2020-08-17T23:53:49.161607Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "20.432816300730064"
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": "calibrator.EstimateCalibrationPointPositions()\ncalibrator.Fit()\ncalibrator.bestError"
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "Note in this run I had to tweak the blur size."
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-17T23:53:56.771683Z",
     "start_time": "2020-08-17T23:53:56.707877Z"
    }
   },
   "outputs": [],
   "source": "calibrator.Save('/E/Driving Data/20200814TZ/Eyetracking/calibration 2.dat')"
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "I now spare you of looking at the other 4 runs, which are the same thing."
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "## Step 3: Get the eyetracking traces, and recenter stimulus frames to be retinotopic.\nThis is where we make the calibrator objects useful. Here we load a saved pupil location file to a new calibrator object, and does things with it.\n\nThere's a separate `Eyetracker` class. One day you won't have to deal with the `EyetrackingCalibrator` class anymore. The `Eyetracker` class handles a lot of meta things, like what if, unlike the driving experinment, you have the calibration video separate from the actual data videos?\n\n#### Recentering frames\nThis process assumes you have individual frames saved out, or can loaded them into a numpy array. We build an `Eyetracker` object here. The important arguments to the constructor are:\n* `calibrator: EyetrackingCalibrator`: this object has estimated a mapping from pupil location to gaze location using the calibraiton video\n* `dataPupilFinder: PupilFinder`: this object has found the Pupil locations in your data video\n* `dataStart: Tuple<int, int, int, int>`: Timestamp of first TTL in the data video\nNote here, I just used the `pupilFinder` of the `calibrator` object, since in my experiment the calibration is in the data video. If it is separate, you will need to build a `PupilFinder` object, preferably of the `TemplatePupilFinder` class because it's faster and less noisy.\n\n`Eyetracker.RecenterFrames` is fairly straightforward. The arguments are a numpy array of frames, and a folder of where to put them. Two other important arguments:\n* `fps: int`: the framerate of your frames, which will most likely be different than the eyetracking videos\n* `scale: float:` how big are these frames relative to the original 1024x768 screen resolution? Motion-energy is often calculated on scaled-down frames since the full resolution is overkill even for the smallest filter.\n\n#### Getting gaze locations\nYou get out the pupil position in video space from a `PupilFinder` object using the `GetTraces` method. Since the videos being before the first TTL, there's a lot of excess fromes. The `PupilFinder` objects' `FindOnsetFrame` method takes a time stamp and gets you the frame index for that time. The 4 arguments are integers for hours, minutes, seconds, and milliseconds.\n\nThen you give those traces to the `TransformToScreenCoordinates` method on the `EyetrackingCalibrator` object to get the corresponding gaze locations."
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-18T01:52:41.400731Z",
     "start_time": "2020-08-18T01:25:29.812532Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0\n29.367180650467834\n1\n20.432816300730064\n2\n29.991116747192002\n3\n20.29448961124729\n4\n28.939680879466003\n5\n26.496719647262243\n"
    }
   ],
   "source": "traces = []\nfor run in range(6):\n    # note that I passed it None so the object doesn't load the eyetracking video.\n    # this saves time and ram\n    calibrator = EyetrackingCalibrator.EyetrackingCalibrator(None, \n                                 starts[run][0], calibrationDelay = 2.0, calibrationDuration = 2.0,\n                                 calibrationPositions = EyetrackingCalibrator.EyetrackingCalibrator.GeneratePoints(),\n                                 calibrationOrder = EyetrackingCalibrator.EyetrackingCalibrator.CalibrationOrder35,\n                                 templates = True)\n    # load a previously saved found pupil locations\n    calibrator.Load('/E/Driving Data/20200814TZ/Eyetracking/calibration {}.dat'.format(run + 1))\n    # we re-fit the mapper because that's quick (and also deterministic)\n    calibrator.EstimateCalibrationPointPositions()\n    calibrator.Fit()\n    print(run)\n    print(calibrator.bestError)\n    \n    # recenter frames\n    frameFiles = glob('/E/Driving Data/20200814TZ/Frames/run {}/*'.format(run + 1))\n    frameFiles.sort()\n    frames = []\n    for frame in range(len(frameFiles)):\n        frames.append(io.imread(frameFiles[frame]))\n    frames = numpy.array(frames)\n    \n    # the Eyetracker object takes the calibrator and does things to video frames with it.\n    eyetracker = Eyetracker.Eyetracker(calibrator = calibrator, dataPupilFinder = calibrator.pupilFinder, \n                                       dataStart = starts[run][0])\n    eyetracker.RecenterFrames(frames, '/E/Driving Data/20200814TZ/Frames Recentered/run {}/'.format(run + 1), fps = 30)\n    \n    # map pupil locations in video space to gaze positions in screen space.\n    firstFrame = calibrator.pupilFinder.FindOnsetFrame(starts[run][0][0], starts[run][0][1], starts[run][0][2], starts[run][0][3])\n    trace = calibrator.pupilFinder.GetTraces()[firstFrame:, :]\n    traces.append(calibrator.TransformToScreenCoordinates(trace = trace))\n    del calibrator"
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": "And now eyetracking videos have been processed."
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": ""
  }
 ],
 "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": "2.7.17"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "source": [],
    "metadata": {
     "collapsed": false
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}