INTRODUCTION - Fundamentals of Three-Dimensional Digital Image Processing - page 11

Standard X-ray images of the human body (meaning not CT, but rather
the previous standard imaging) is not quite so precise, yet it too has its own
depth of information. In fact, by examining several X-ray images taken at
different angles, one can to some extent recreate a 3D image showing the
internal structure of organs and blood vessels.
(f) Animated images
A true 3D image is stored in a 3D array and moreover has three axes ( i, j, k ),
indicating position in space. In other cases, a time series of 2D images (in
other words, an animation) may be stored in a 3D array and treated as a
single 3D image. In such cases these objects are treated as true 3D images
having rectangular image elements (in most cases, the direction of the time
axis and that of the other axes will represent different physical meanings and
resolutions), but generally they are not referred to as 3D images. Instead they
are often called spatiotemporal images. (In other cases, images may show
foreground and background relationships between objects due to occlusion, or
as discussed above an otherwise 2D image may carry some 3D information.
Such images are often called 2.5D images.) Video and television images are
trivial examples. Some processing, for example, tracking and measuring the
movement of specific objects in an image or editing support, is effectively
performed when the image is treated as a 3D one. There has been interesting
research in recent years related to creating a virtual environment (including
temporal changes) within a computer in real time, based on a 3D space created
using images from multiple TV cameras, and furthermore adding 3D image
data for objects that do not actually exist. This is referred to as mixed reality
[Ohta99].
(g) Variable focus microscopy images
By using a device that can view a single object from a fixed angle, making
fine adjustments to the position of the focus surface (that object surface on
which the focus is aimed), and taking a series of photographs after moving the
focus surface by tiny amounts, one can create a composite image that records
surface irregularities in the direction of the optic axis. This allows for distance
measurements between points on the visible surface of the object along the
primary axis of the photographic lens. A 3D image of those surface irregular-
ities then can be created. If the object of study is transparent, then observa-
tion of its internal structure is also possible. A well-known example of such a
device is the confocal laser microscope. Examples of applications include mi-
croscopy of tissue samples from living specimens and examination of surface
features on VLSI devices [Ichikawa94, Ichikawa95, Ichikawa96, Vitria96]. See
also [Geissler99].