Plot the Distribution of Tracking Confidence

Renders a check_confidence() result (from the anicheck package) as a horizontal violin per keypoint of its tracking confidence (keypoints on the y axis, confidence on the x), with a point at the median and a line spanning the inter-quartile range. A violin sitting far to the right and tight is a reliably tracked keypoint; a left-shifted body or a long low tail flags one the tracker was often unsure about. With several individuals, each gets its own panel.

Usage

# S3 method for class 'anivis_check_confidence'
plot(x, ..., clip = 0.02, mode = c("light", "dark"))

Arguments

x

A check_confidence object (from the anicheck package).

...

Additional arguments (currently unused).

clip

Cut each violin where its density falls below this fraction of its peak, so thin tails and the neck bridging a bimodal distribution are removed and the violin shows only where data actually is. Default 0.02; set 0 to keep the full density.

mode

Either "light" (default) or "dark"; passed to theme_imputets().

Value

A ggplot object.

Details

The violins are drawn from the kernel-density grid stored in the check object (via geom_polygon), so no raw values are needed. Styling matches the other check plots (theme_imputets(), horizontal-only gridlines). The plot is built from an intermediate frame of class anivis_check_confidence_data produced by as_plot_data() — the staging step that mirrors data_plot() in see.

See also

as_plot_data()