Morphometric Data - An Invariant Approach to Statistical Analysis of Shapes - page 32

three-dimensional configuration by simply translating and rotating
the common landmarks is recorded in one view (e.g., landmarks from
the right side of an object) to match exactly the coordinates of the same
common points collected from the other view (e.g., left side). Computer
programs can facilitate this process. Since the relative location of the
common landmarks are identical in the first and second set (barring
significant measurement error), the coordinates of the common land-
marks can be matched exactly, and the relative location of the
remaining landmarks collected from the two sides are brought into
this single coordinate system.
2.3.3 Landmarks from serial sections or captured images
Embryonic, fetal, and other histological and anatomical specimens are
often serially sectioned in order to reveal aspects of internal structure.
Historically, reconstructions of these sections have been done to reveal
the three-dimensional structure of internal features. Two-dimensional
data can be collected from each slice, and when special precautions are
taken to enable registration of the sections to one another, a surrogate
coordinate of the third dimension can be added using information from
the relative location and thickness of the particular slice. Working with
serially sectioned specimens introduces new avenues for potential
error. First, error can be introduced by the methods of registration and
alignment used in reconstruction of the sectioned specimen.
Additionally, Lozanoff (1992) points out that tissues may distort or
shrink if subjected to fixation and dehydration. Distorted images may
result even when tissues are normal if the optical properties of the
video camera or the configuration of the video board are flawed
(Lozanoff, 1992). These problems may seem overwhelming to the
untrained histomorphologist who wants to obtain reliable landmark
data from histological sections. Lozanoff (1992) has demonstrated an
alternative approach that involves the recording of images of sectioned
surfaces prior to microtome sectioning. He provides the accuracy and
precision of measurements taken from computerized three-dimension-
al reconstructed models made from these images.