Information Technology Reference
In-Depth Information
compose musical symbols the same way as on a
sheet of paper. The gestures are then as classical
as possible, as presented in the sixth column of
Table 1. Figure 4 presents two screenshots of the
editor: on the left, the user draws a beam between
two quarter-notes, and on the right, this drawing
is interpreted and retranscribed neatly.
Screenshots of the pen-based musical score
editor developed with the presented methodol-
ogy: on the left, the user draws a beam between
two quarter-notes, and on the right it is replaced
by its neatly retranscribed symbol thanks to an
eager interpretation process.
to the other elements of the document; this
is its main originality. This formalism can
be applied to different notation domains and
composition conventions, in order to model
how to eagerly interpret a very large panel
of hand-drawn structured documents. The
interpretation rules written for one domain
according to this formalism are externalized
from the system: they are easily modifiable.
They just need to be transformed by the
generic compiler of the system. This formal-
ism is presented more in detail in the first
subsection.
A hand-drawn stroke analyzer (2), which
exploits the knowledge modeled by the
formalism to eagerly interpret hand-drawn
structured documents. It is able to call the
pertinent hand-drawn shape recognizers
(for example clef recognizer, accidental
recognizer, etc.) depending on the structural
context of the stroke, and then update the
document contexts that will help recogniz-
ing the following strokes. This analyzer is
generic: it does not need to be adapted to each
specific domain. This analyzer is presented
more in detail in the second subsection.
a generIc approach for
hand-draWn structured
document composItIon and
edItIng
In this section, we present a new generic approach
that aims at interpreting hand-drawn structured
documents from various domains and with differ-
ent constraints, such as musical scores, but also
diagrams, plans, electronic figures, and so forth.
The architecture of the system is illustrated in
Figure 5. It is based on a framework constituted
of three main components:
A set of graphical functions and pen-
based editing functions (3), which can be
exploited by any pen-based system; such
functions deal with the user interface and
the display of the document. They exploit
graphical information, such as images of
the neatly retranscribed symbols, which are
externalized because they are dependent on
the domain. These functions also deal with
pen-based interaction and editing: the user
can for instance select graphical elements,
move them to another part of the document,
erase them, and so forth. It is also possible
to undo the last element, zoom in or out,
save or load a document. All these actions
can be done just using the pen. These func-
tions are presented more in detail in the last
subsection.
A formalism for eager interpretation of
hand-drawn structured documents (1),
which models how to eagerly interpret the
elements of hand-drawn structured docu-
ments from a given domain. It permits to
write interpretation rules that represent
composition conventions (for instance
chronological information: it models which
element can or must be drawn before another
one) and physical information, such as the
spatial structure of the document. Finally,
they model the driving of the use of hand-
drawn shape recognizers depending on the
document structural context of a stroke,
that is depending on its location in relation
Search WWH ::




Custom Search