Skip to content

Gaze Meta Data

Swartz Center for Computational Neuroscience edited this page Jul 21, 2015 · 3 revisions

This page specifies the meta-data for gaze-parameter recordings (e.g., from an eye tracker).

If a stream has the content-type Gaze, we recommend that meta-data about the stream adheres to the structure and naming laid out in the following. While all meta-data is optional, we recommend that any stream should describe at least for each channel the label, eye, type and unit parameters.

<channels>           # specification of the channel layout 
  <channel>          # information about a single channel (repeated for each)
    <label>          # label of the channel 
    <eye>            # which eye the channel is referring to (can be left, right, or both)
    <type>           # type of data in this channel, can be an of the following values:
                     # ** ScreenX, ScreenY: screen coordinates of the gaze cursor (can also refer to a scene image), usually in pixels,
                     # ** DirectionX,DirectionY,DirectionZ: 3d gaze vector in some coordinate system
                     # ** PositionX,PositionY,PositionZ: 3d position of the eye center in some coordinate system
                     # ** IntersectionX, IntersectionY, IntersectionZ: 2d or 3d position of the intersection point with a plane (in some coordinate system)
                     # ** HeadX, HeadY, HeadZ: 3d location of the head center in some coordinate system,
                     # ** PupilX, PupilY, PupilZ: 2d or 3d location of the pupil center in some coordinate system,
                     # ** ReflexX, ReflexY, ReflexZ: 2d or 3d location of the illuminator's reflection point in some coordinate system,
                     # ** Radius or Diameter: the overall pupil radius or diameter (usually in mm or pixels),
                     # ** RadiusX,RadiusY: horizontal and vertical pupil radius
                     # ** DiameterX,DiameterY: horizontal and vertical pupil diameter
                     # ** Confidence for confidence information (preferred unit: normalized)
                     # ** FrameNumber: frame number that the parameters were calculated from
                     # * PlaneNumber or ObjectId: number or identifier of the object that was intersected by the gaze vector
    <unit>           # measurement unit (e.g., pixels, mm, normalized)
    <coordinate_system> # coordinate system of the respective parameter, can be world-space, object-space, camera-space, or image-space
  </channel> 
</channels> 

<acquisition>        # information about the acquisition system 
  <manufacturer>     # manufacturer of the device
  <model>            # model name of the device
  <serial_number>    # serial number of the device
</acquisition>