Version 1.2.0

• Rework median filtering and correlation functions to use OpenCV instead of
  scipy for performance improvements.
• Fix seed point finding method to use template matching to improve point finding
  with the default bounding boxes.
• Add keyword arguments to filters for candidate points.
• Expose additional input parameters: average_iris_intensity, clip_pupil_values,
  and max_eccentricity.
• Add constraints to EllipseFitter, preventing any ellipse axis longer than the
  index ray length as well as limiting eccentricity to below max_eccentricity.
• Use the keyword arguments for candidate point filters to filter rays where a
  the baseline intensity is out of pupil limits if clip_pupil_values is set.
• Add plot of average pupil intensity to QC output to check behavior of adaptive
  pupil tracking.
• Add plot of best fit error to QC output.
• Add UI for testing configuration parameters and generating input jsons.