QuickDraw: Improving Drawing Experience for Geometric Diagrams
CHI 2012, May 2012, Austin, Texas, USA
Salman Cheema
· University of Central Florida
· Pursuing a Doctoral degree in Computer Science
· Area of primary interest is Computer Graphics.
· http://www.eecs.ucf.edu/isuelab/people/salman.php
Sumit Gulwani
· Microsoft Research
· Research interests are in the cross-disciplinary application areas of automating end-user programming
· http://research.microsoft.com/en-us/um/people/sumitg/
Joseph J. LaViola Jr.
· University of Central Florida
· “interests include pen-based computing, 3D user interfaces for games, human motion estimation, virtual reality, and interactive computer graphics”
· http://www.eecs.ucf.edu/~jjl/
Summary
Students and teachers in scientific disciplines often have
to draw very precise diagrams. “Our research goal is to investigate the use
of sketch-based interfaces for drawing precise diagrams since they can provide
significant benefits over existing tools.” QuickDraw is a prototype
sketch-based diagram drawing tool that allows the user to sketch and beautify a
given diagram.
“Sketch recognition is a hard problem due to the complexity
and variation in notation in different disciplines.” Techniques for sketch beautification have
also been explored by several researchers.
QuickDraw infers a user’s intent by recognizing constraints relating
sketched primitive components and uses this information to beautify the entire
sketch instead of just individual components.
In its current form QuickDraw can recognize and beautify
diagrams containing line segments and circles.
Recognition in QuickDraw is a two-step process: the sketch is parsed
into diagram (basic) components and the basic components are determined to be
line segments or circles. The inference
subsystem infers the intended constraints between recognized components. The beautification algorithm has two
interesting characteristics: robustness and interactive support. Users can sketch a drawing in two ways:
sketch the entire drawing and then trigger the recognize button or sketch
incrementally, hitting the recognize button after each step.
 |
| Figure 1. |
The users were to compare QuickDraw with traditional WIMP
based drawing tools. The researchers recruited a variety of people for the
study. Three levels of difficulty were
tested but users were not told the difficulty of the drawing being done. Figure
1 shows some of the figures tested by the users. The users filled out a pre and post
questionnaire and were given practice diagrams to familiarize themselves with
the programs.
The overall findings were: with easy difficulty, there was
only a difference between QuickDraw and PowerPoint and with medium and hard
difficulties, the only NON-different program was Geometry Expressions. Some of the factors affecting completion time
are: the order in which a tool is used, the drawing strategy: perfection or
not, quick or careful, triggering after the whole sketch is completed or after
each part, and incorrectly inferred constraints.
The first section asked participants to rate each
diagramming tool on a number of qualitative metrics. The second section contained questions
related only to QuickDraw. There were
significant differences for overall reaction, perceived drawing performance,
and the ability of each tool to enable easy diagram drawing. QuickDraw was better than PowerPoint, but not
than the other options.
Performance is hampered by three major factors: high failure
rate, lack of adequate editing capabilities, and drawing diagrams incrementally
inflates the completion time. In the future,
the authors would like to improve the editing abilities in QuickDraw, the inference
constraints by using classifiers over a variety of features and to extend the
constraint language to refer to virtual components.
Related Work
1.
“Sketch based interfaces: early processing for sketch understanding” –
This paper covers the topic of transforming a freehand sketch into a computer
model.
2.
“Sketched symbol recognition using Zernike moments” –
This paper covers recognition for hand drawn symbols in an online method.
3.
“A sketch-based interface for iterative design and analysis of 3D
objects” - This paper describes a program that works in conjunction with CAD to
make freehand sketch based engineering design.
4.
“Free-sketch recognition: putting the chi in
sketching” – This paper covers the difference between gesture based and free
sketch recognition techniques.
5.
“Sketch-based modeling: A survey” – This paper
covers the ideas behind mapping a 2D sketch to a 3D model.
6.
“Sim-U-Sketch: a sketch-based interface for
SimuLink” – This paper details an experimental sketch-based interface to
construct correct SimuLink models.
7.
“A Parsing Technique for Sketch Recognition Systems” – This paper covers “a
framework for modeling sketch languages and for generating parsers to recognize
them.”
8.
“Towards a computational model of sketching” –
This paper covers the need for computers to be as effective at sketching and
recognizing sketches as humans are. It
also poses several guidelines for determining when this happens.
9.
“Recognition of freehand sketches using mean
shift” – This paper proposes a “mean shift” method of sketch recognition. This method does not require any specific
knowledge, and thus can be applied to anything.
10.
“Beautifying sketching-based design tool
content: issues and experiences” – This paper covers a significant number of
problems posed by the beautification of a sketch. Several examples are given to explain better.
As is clearly obvious from the above listed papers, this topic is
not novel. The only novel thing about it
is using sketch recognition for strictly geometric diagrams.
Evaluation
The work was evaluated in a variety of ways. The users were evaluated with pre and post questionnaire
allowing for quantitative, subjective results.
These questions were based on how the user thought the study was
organized and what the user thought about the QuickDraw program. The researchers were also recording two
numbers which were quantitative, objective (the time and number of edits). These numbers allow the researchers to draw
conclusions that the questionnaires may not.
Because of the multiple types of evaluation, the data covers a large
amount of data and is more likely to be reproduced.
Discussion
This work is not entirely novel, it is being researched in a
variety of locations because of this it is interesting topic. The variety of people researching this topic
means there are a lot of possible solutions being created. Overall, the research was an interesting read
if nothing else. The evaluation done was
highly effective. It allowed for users
to be subjective in their results but still have the data be numerical. The variety in users could have been greater;
personally I think that you need more than 17 people to conduct a study. Also, women and men use programs differently
so I would have tried to include more women in the study.
No comments:
Post a Comment