30 2. FIVE STORIES TO A MODEL OF VIDEO STRUCTURE
bundle. e frame transects the signal bundle and reveals a matrix of strand values. e frame sets
the vertical and horizontal bounds of the matrix; the cells of the matrix are pixels (picture elements).
e frame is the constant, regular, and addressable unit of measure within a (MID). It is a
sampling mechanism which provides a reference point within the length of the document and it
provides a set of X,Y coordinates for any individual sample point within the document. is makes
possible comparisons within individual frames and among frames. e Z-axis is the third compo-
nent, allowing measurements across frames (essentially, time). It should be noted here that in the
days of analog lmmaking, while the frame was the unit for making editing decisions, elements
within the frame could still be discussed and used—making the cut at the point where the hand
reaches just above the horizon; waiting until about one third of the frame showed the pink of the
sunset sky; cutting where the movement of the horses foot at the end of one sequence just matches
the location and movement of the seagull in the next.
e pixel is the smallest addressable unit within an individual frame. Each pixel represents
the value of a strand of the signal bundle at a particular time. e value of each strand of the signal
bundle can be viewed relative to the values of all other strands in the bundle at a particular plane
(time), as well as to the values in the same strand and other strands at other frame planes.
e representations present patterns of changes in the relationships of pixel values, rather
than extractions based on mechanical demarcations (e.g., shot or cademe or word). Each pixel can
be addressed with X, Y, Z coordinates. Precise analysis, comparison, and display of MID elements
can be made by precise numeric and formulaic description of location (X, Y) and extent (Z). Such
an approach removes much of the ambiguity of earlier descriptive terminology. It also allows for
tracking one document attribute across boundaries of other attributes (e.g., music continuing even
when a “shot” has ended).
2.3.3 ANALYSIS ENVIRONMENT
Extraction of key frames, in a manner analogous to the extraction of key words in a verbal text,
is accomplished by the identication of points of signicant change of spatial and temporal data.
is is very similar to the way automatic indexing of word-based documents identies keywords by
looking for points of change and signicant frequency of occurrence. is system makes use of the
numerical precision of the digital environment and the processing capabilities that can be applied
to the numerical data to measure temporal aspects, size of an object or portion of an image, and the
coordinates of an image segment within the eld.
During the extraction process, the program tools are superimposed over the images being
analyzed. e software tracks attributes of user-identied objects:
• point at which object appears,
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.