AutoMovie API
    Preparing search index...

    Namespace AutoMovieHumanReferenceOutlines

    Outline data between the published measures that size the view parts: private-reference knot tables and the separately cited public upper-lid Bézier control points, SHIN_UPPER_LID.

    Estimated from the private reference set (f1, f2, m1 and m2, aggregated as normalized ratios, .references/curves); no published source carries these shapes. Each table names the reading that produced it under test/src/features/human/estimate/, its normalization, and its consumer; each row is a station share and the mean of each column over the people the station read on, with its SD over them. A cell is kept when read on at least two people with an SD of at most 0.25 of its normalizing measure (the per-stature mass tables: at most its mean's magnitude); a station whose reading fell apart across the people, as where the pack's arcs miss a side, leaves that cell out and the column reads through from its neighbours. A table whose rows are empty would wait for its reading's run, and its reader reports that rather than inventing a shape.

    Scaling: the private tables contain shares of the person's own measures; stations are shares between landmarks. SHIN_UPPER_LID instead retains its published photo-scale millimetres and pupil datum for an explicit mapping.

    BREAST
    BREAST_FOOTPRINT
    EAR
    FACE
    FISSURE
    FOREARM
    GLUTEAL
    GLUTEAL_FOOTPRINT
    LIPS
    SHANK_SPLIT
    SHIN_UPPER_LID
    THIGH_SPLIT
    TRUNK
    UPPER_ARM
    at