|
|
|
Related Papers:
- T. Hammond and R. Davis, "Tahuti: A
Geometrical Sketch Recognition System for UML Class Diagrams", AAAI Spring Symposium on
Sketch Understanding, March 2002, pp.59-68.
Summary: Tahuti is a tool that recognizes UML diagrams based on how they look like rather
than how they were drawn. It recognizes eight UML objects: general class, interface class,
inheritance association, aggregation association, dependency association, interface association,
text and unrecognized strokes. Tahuti also recognizes two types of commands: delete and move.
The recognition process consists of six stages:
- Preprocessing: In this layer a stroke is processed to fit in one of the following categories:
1) ellipse, 2) line, 3) polyline (collection of line segments), and 4) complex shape
(collecttion of line segments and bezier curves.)
- Selection: After a stroke is preprocessed, it is then combined with zero or more unrecognized
strokes. The system combines the last nine unrecognized strokes that are within 10 pixels
of each other and it creates all possible stroke collections.
- Recognition: In this layer all stroke collections are examined to see if they can be recognized
as a UML object or an editing command. The final decision of how to interpret a collection
is defered until the next stage.
- Identification: In this stage the system classifies stroke collections. The Move command has
a priority over any other possible interpretation. Also collections with larger number of
strokes has priority.
In the future Tahuti will be connected to a CASE tool like Rational Rose to enable the
programmer to use the functionalities provided by these tools.
- T. Hammond and R. Davis, "
LADDER: A Language to Describe Drawing, Display, and Editing in Sketch Recognition",
Proceedings of the 2003 Internaltional Joint Conference on Artificial Intelligence (IJCAI),
August 2003.
Summary: Most sketch recognition systems are domain-specific. The authors are proposing
using a domain-independent recognizer that uses a domain specific sketch grammar. This way
users can create new sketch interfaces by writing a sketch grammar for the domain they are working
with. LADDER's language consists of predefined shapes, constraints, editing behaviors, and display
methods.
- A. Caetano, N. Goulart, M. Fonseca and J. Jorge, "
JavaSketchIt: Issues in Sketching the Look of User Interfaces", AAAI Spring Symposium on
Sketch Understanding, March 2002, pp.9-14.
Summary: JavaSketchIt recognizes few simple shapes (line, circle, rectangle, etc.) using
CALI recognizer. It then recognizes more complex shapes as a combination of those simple ones.
It uses Fuzzy Relational
Grammars to specify the attributes of a complex object.
- "A visual language parsing based on automata" Sangwook Kim; Mansoo Kim; Youngmee Shin Source:
Journal of the Korea Information Science Society, v 21, n 2, Feb. 1994, p 393-406 Language:
Korean ISSN: 0258-9125 CODEN: HJKHDC
- "Visual gesture recognition" Davis, J. (Univ of Central Florida); Shah, M. Source: IEE
Proceedings: Vision, Image and Signal Processing, v 141, n 2, Apr, 1994, p 101-106 ISSN:
1350-245X CODEN: IVIPEK
Related Books:
Related Applications:
- CALI:A Software Library for Calligraphic
Interfaces
- SATIN: A Toolkit for Informal Ink-based
Applications
- Tahuti: A Geometrical Sketch
Recognition System for UML Class Diagrams
- Ideogramic UML: UML tool for the Tablet PC
- Tablet UML: Pen-based diagramming tool
Related Conferences:
- CHI (Computer-Human Interaction)
- CSCW in Design (Computer Supported Cooperative Work in
Design)
- DIS (Symposium on Designing
Interactive Systems)
- HCI International (International Conference on
Human-Computer Interaction)
- IASTED (International
Association of Science and Technology for Development)
- IUI (International Conference on Intelligent User Interfaces)
- UIST (Symposium on User Interface Software and Technology)
- VL/HCC (IEEE Symposium on Visual Languages
and Human-Centric Computing)
Related Journals:
Other links:
|
