Conference PaperPDF Available

Freehand sketch system for 3D geometric modeling

Authors:

Abstract

The paper introduces a new method to deal with sketches for inputting geometric models at a workstation. The sketches are drawn on a CRT screen through a stylus pencil and a tablet by a designer at an early stage in the design procedure. The designer can keep the drawing stylus at a workstation as the same manner as using a pen on paper. The system `Sketch Interpreter' can correct geometric models in a computer even though input sketches are geometrically distorted. This system can create 3D geometric models in a computer even though input data of 2D sketches are drawn by freehand. The system is constructed in terms of three characteristic procedures on a workstation with a pointing device; (1) to recognize hand sketches on the screen. Freehand lines are drawn by pen operations from which a computer can construct a 3D geometric model. The first operation computes mathematical parameters for the projective transformation; (2) to construct additional geometric models by inputting more sketches drawn perspectively; and (3) to redraw modified geometric models replacing the sketch lines. The interactive methods are suitable for realistically constructing any geometric shapes that a designer imagines. Created data are transferred to an advanced 3D-CAD system. The system is applied as a front-end processor of design practice in fields such as office equipment
Freehand Sketch System
for
Koichi
MATSUDA, Satoru SUGISHITA,
Zheng
XU,
Kunio
KONDO,
Hisashi
SAT
and
Shizuo
SHIMADA
Dept.
of
Information and Computer Sciences, SAITAMA University
Shimo-okubo
255, URAWA, SAITAMR
338,
JAPAN
{matsuda,kondo}@ke.ics.saitama-u.ac.jp
Abstract
This paper introduces a new method to deal with
sketches for inputting geometric models at a worksta-
tion. The sketches are drawn on a
CRT
screen through
a stylus pencil and a tablet by a designer at an early
stage of design procedures. He can keep their drawing
stylus at a workstation as the same manner as using a
pen on paper. The system ‘Sketch Interpreter’ can cor-
rect geometric models
in
a computer even though input
sketches are geometrically distorted. This system can
create
30
geometric models
in
a computer even though
input
data of
2D
sketches are drawn by freehand. The
system
is
constructed
in
terms of three characteristic
procedures on a workstation with a pointing device
;
(1)
to recognize hand sketches on the screen. Free-
hand bines are drawn by pen operations
from
which a
computer can construct a
30
geometric model. The
first operation computes mathematical parameters for
the projective transformation;
(2)
to consiruct addi-
tional geometric models by
inputting
more sketches
drawn perspectively; and
(3)
to
redraw modified geo-
metric models replacing the sketch lines. Our interac-
tive methods suit to realistically construct any geomet-
ric shapes a designer imagines. Data created are trans-
ferred to an advanced
3D-CAD
system. Our system is
applied as a front-end-processor of design practice such
as for ofice equapments.
1.
Introduction
At early stages of shape design,
a
designer always
tries to make his new ideas into sketches as soon as
possible. Ullman
[l]
argued that sketching is an essen-
tial activity for creative design because it allows
a
de-
signer to think aloud and to evaluate new ideas quickly.
It also assists the designer’s short term memory and
communication with other people. But current
CAD
systems do not
suit
for trying out
a
variety
of
ideas
by
thumbnail sketching. Binh Pham
[2]
mentioned
CA
still falls short of designer’s expectation, they stili re-
main to be seen more as
a
drafting tool than
a
desi
tool. Designers cannot u5e
or
at
best feel very
uficom-
fortable to use these systems
for
the early stages of
design. In current
CAD
system, specifications for geo-
metric input and constraints need to be very specific,
and unconventient to make free-form surfaces, hence
they are only useful
in
the final stages
of
design when
a designer knows exactly what the object should look
like.
In engineering design, training
is
required
for
a
de-
signer to imagine
a
3D
object that does not exist yet,
and
tG
transform it into
a
2D
sketch before making en-
gineering drawings. At the stage to imagine
a
shape,
a
designer makes brain storming by himself
for
drawing
sketches. However, at the design of engineering shapes,
sketches are drawn at an early stage of design proce-
dures, then
3D
models must be constructed
in
the
fol-
lowing many stages. Most advanced
JD-CAD
systems
demand accurate dimensions of shape to
a
designer,
and will not accept vague information as shown
in
the
sketches. The shape of
a
target object
is,
therefore, to
be modified frequently during stages
of
engineerin
sign,
so
that it is desirable to shorten the times
for
such
repeated stages. The aims of our study are to build
a
front-end processor which creates
3D
mode!s from de-
signer’s idea and to transfer the data for an advanced
3D-CAD
system.
Typical techniques to recognize
a
3D
object
from
a
2D drawing have been using the perspective theory
[3][4][5][6]
or
the orthographic view theory [7][8].
‘In
the
perspective theory,
3D
models can be created from
2D
drawings which are taken into
a
computer by
a,
scanner.
For
an instance,
a
lot of efforts have been paid to
Skq-
plify the processes to generate
3D
models
by
scanning
sketches. Akeo[6] allows users to scan sketches vhich
includes perspective vanishing lines and
3D
cross
sec-
55
0-8186-7867-4/96 $10.00
0
1997 IEEE
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
tions into the computer. The scanned data is projected
onto the 3D mark-up to complete the process with var-
ious editing.
In general,
a
2D view shows less information about
sizes of
a
3D object,
so
that designers have to fulfill ad-
ditional information such as depth sizes. We do not use
the orthographic view theory in this paper because our
aims deal with sketches mostly drawn perspectively.
Our Sketch Interpreter is an interactive system based
on the perspective theory. Our system is unique, since
a
sketch drawn on
a
CRT
screen is directly recognized
by
a
computer as
a
3D object. Designers can draw
sketches interactively by Sketch Interpreter as the same
manner as drawing sketches on paper. Sketch Inter-
preter records histories of designer's operation,
so
that
the lack of information can be recovered by the addi-
tional data with historical information.
Sketch Interpreter helps designers of shape design
especially for small industrial parts, office equipments,
wooden furniture in living house, etc. Textile and color
design to 3D models are not discussed in this paper.
2.
Overview
of
Sketch Interpreter
2.1.
Systems Concept
Engineering design is generally composed
of
six
stages of procedures (Fig.
1).
At every stage, the shape
of
a
target object is inspected from various aspects such
as artistic design, functional requirements, and
so
on.
When these are not satisfied,
a
modified shape is tested
from the preceding steps. Trial and correction proce-
dures take usually
a
lot of hours,
so
that it is desirable
to shorten the repeated steps. At the stage to imag-
ine
a
shape,
a
designer draws any possible ideas onto
sketches.
lil
Figure
1.
Design processes
Figure 2 exemplifies procedures of design ideas
which are usually drawn by
a
designer on paper. Figure
2-a
shows
a
target object by outlines of parallelepiped
at
the beginning of sketches drawings. More details are
then added over the rectangular object (Fig. 2-b), and
finally
a
clean copy is completed(Fig. 2-c).
Figure
2.
Hand
sketch procedures
on
paper
A designer draws rough sketches on the tablet as the
same manners as on paper, but the figures are drawn on
the graphics screen of
a
computer. Our system adopts
nearly the same manners as pencil drawings on paper,
since
a
designers is not always acquainted with com-
puter handling.
A
designer draws
a
perspective shape
of some possible rectangular parallelepipeds onto the
screen with freehand lines. A computer adjusts the
freehand lines into straight edges. Vertices are decided
as starting positions of relevant edges. Two
or
three
vanishing points in the perspective theory are reason-
ably proposed by the computer, and
a
geometrically
correct perspective view is redrawn on the screen. At
this moment,
a
view position is decided as representing
an eye of the designer. We call this perspective shape
as 'Basic Shape' which plays an important role in
our
system.
A designer then draws additional freehand lines over
the basic shape to make more details
as
though he
complete
a
picture on
a
canvas. The freehand lines
may indicate cutting lines to the model and the cut-
ting operation is carried out by pointing the part to
be eliminated. Models on the screen can be inspected
perspectively changing view positions(Fig.
3).
2.2.
Steps
of
Drawing Stage
Based on the perspective theory the system can rec-
ognize
a
2D sketch and change it into
a
3D geometric
model even though the input sketch data are uncer-
tain. The designer can also manage adding and cut-
ting operations on the basic shape, and make rounding
operation to the edges and corners. Figure
4
shows the
steps of procedures of
our
system.
At the first step,
a
designer draws freehand lines on
the screen through the tablet(Fig. 4-a). The freehand
lines are
a
series of positions which become
a
chain
56
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
a
j
draw an outline add details
(dr
Lch)
I
acomputer
I
t
(calculate ID-data)
I
restore
to
a
3D
objmt
*
modify
3D
object
Figure
3.
Computer
assisted
sketch proce-
dures
of pixels, but are not yet straight lines on the screen.
Five procedures are designed to correct the shape: line
recognition, unifying nodes, numbering on vertices, ad-
justing edge direction, and restoring to
a
basic shape
in
a
3D
geometric model. A computer calculates these
freehand lines to become edges of
a
3D
object and com-
pletes
a
‘Basic Shape’ (Fig. 4-b, c, d, e). As the same
manners, the designer draws more additional freehand
lines over the basic shape to add
or
to cut
off
parts
from the basic shape and makes rounding operation on
edges and corners(Fig. 4f,
g).
a.
Ptxel drawmg Line drawing Basic Shape Rectifying
twists
Restoring
to
3D-object Detailed drawings Rounding operation
Figure
4.
Steps
in
Sketch
Interpreter
3.
Thinning Algorithm for Interactive
Line Drawing
This section introduces the methods to fix an edge
line from freehand sketch lines. Practically
a
proposed
edge line is drawn by many overlapped lines by
a
de-
signer. He repeatedly draws such lines roughly the first
and more concretely the later.
A
freehand line drawn
later and later becomes preferable to be the proposed
edge line. We edit therefore an edge line from freehand
lines considering these time sequential input. The algo-
rithm falls into three categories
:
segment composing,
clipping, and segment joining.
3.1.
Segment Composing
When
a
designer proposes an edge line with
a
pencil
on paper, he composes the line from several stroke of
freehand lines. A segment is
a
straight line bundled
from
a
or
more freehand lines which are drawn during
a
short time
of
period. Another segment can be overlaid
to the preceding segment, and then two segments are
composed into
a
renewal segment. A segment has two
end points, vertices, which are the beginning and end
points of
a
freehand line. The freehand line is tested in
its shape whether it is nearly
a
straight line. When it
shows
a
curved line within
a
strip, it is assumed that
two
or
more arcs are connected to become
a
folded
line composed of two
or
more segments. The ratio
of
width to length of the strip is decided
as
about
0.1
experimentally. In order to test two segments, each
segment is identified by the coordinates
of
its middle
point and an angle
of
inclination from the horizontal
line. Four sorts of tests are carried out to compare
two segments whether they are independent
or
they
must be edited into one, that is, angles, distance, time
period, and length(Fig.
5).
Figure
5.
Composing
stroke
segments
(1)
test about angles
In order to compare two segments by angle of in-
clination, the relative angles
cr
and
/3
are calculated
as
shown in Fig.
6.
Accounting the magnitude of angles
cr
and
/3,
two segments will be unified into one segment
or
prolonged into
a
longer segment.
Figure
6.
Angle
difference
of
two
stroke
segments
57
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
(2)
test about distance
Distance of the middle point of
a
segment to an-
other segment line is calculated(Fig.
7).
When the
distance is bigger than
a
threshold, two segments are
independent.
Figure
7.
Distance between two segments
(3)
test about time sequence
We observed experiments how
a
designer builds his
sketches practically, and found that the later drawn
lines are more adopted then the former lines.
In
order
to synthesize two segments into one, the new segment
is
calculated as an algebraical mean with weight ratio
X
and
py
where the ratio is experimentally decided as
X:p=l:l
(4)
test about length
When two segments are edited into
a
new segment,
they are compared by their length, and the overall
!ongest length
is
adopted for the new segment.
.2.
Clipping and Joining
After many segments have been drawn indepen-
dently on the screen, every possible vertex has two
or
more segments which shall be edges joining
at
a
po-
sition. These segments may not meet
or
cross
at
the
position(Fig. 8-a). Intersected points are calculated
among the segments. When two
or
more intersected
points exist
on
a
segment, the far lying point is adopted
as
a
vertex to this segment as well as to the opponent
segment(Fig. 8-b). There may be still two
or
more
intersected points around
a
proposed vertex which is
then decided
as
a
gravity center of such points(Fig
.8-
c). After these procedures against all the segments,
vertices and edges are then decided.
As
described in the article
2.2,
a
basic shape
is
drawn
as
a
perspective view of
a
parallelepiped
for
a container
to
a
3D
model. The figure is
a
2D
image look like
Figure
8.
Clipping and Joining
(a)
sketches
(b)
Line composing
(e) Clipping
(d)
Result
Figure
9.
Thinning example
a
map. The data of the image must be then recon-
structed to have
3D
properties. Sketch Interpreter has
functions for this purpose by numbering to vertices,
adjusting edge direction, and decision
of
a
view point
for
perspective transformation.
4.1.
Numbering
to
vertices
A
3D
parallelepiped has 8 vertices, however the basic
shape drawn perspectively shows
at
most seven vertices
on the screen. Labels to these vertices are decided be-
forehand as shown in
Fig.
10.
A
designer has composed
edges by segments for the basic shape in an arbitrary
order onto the screen. We prepare
a
fixed memory area
to hold the data of edges and vertices. At most three
surfaces are visible
for
a
view
of
a
parallelepiped.
A
loop is considered to trace along edges around
a
sur-
face in reversal clockwise turn. A
or
more common
edges are found to bound two surfaces.
A
vertical edge
of such common edges gives the key for numbering to
vertices.
58
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
4.2.
Adjusting edge direction
A designer at first makes his sketch regardless
of
ge-
ometric accuracy to keep the perspective theory. This
causes that extended edges of
a
parallelepiped may not
meet at
a
fixed vanishing point.
A
proposed vanishing
point
is
decided along one of the common edges as the
mean position
of
other intersected points (Fig
.lo).
In
addition, two vanishing points are set on the horizontal
line. After a
or
two vanishing points are decided, all
edges except the common edges are slightly rotated to
aim such vanishing points.
decide vanislung points rotate edges
Figure
10.
Adjusting edge direction
4.3.
Decision
of
a
view point
A perspective view
is
graphically displayed on the
theoretical projective plane through
a
point which
stands with
a
distance
f
from the plane. Practically
the perspective view
is
obtained by
a
camera,
so
that
we
can suppose an imaginary camera with focus length
f,
the problem to decide
a
view points is to deduce the
focus length
f
and the center of photograph. The view
point is the position
of
lens and stands spatially with
the distance
f
from the perspective view. The view
point is obtained by graphics geometry
as
shown in
Fig.
11
under the conditions that every surface is the
right angle rectangle in
3D
space[4].
4.4.
Restoring
3-D
coordinate
system
We can extract the geometric solid model from the
2D
sketch by the method introduced by Kondo
[4][9].
A
transformation matrix is deduced from the relations.
The world coordinates
of
all the vertices can be then
induced from the data of graphic image.
\is
Figlare
11.
Decision
of
a
view
point
lar parallelepiped(Fig.
12-2).
The system calculates
these freehand lines to edges
as
2D
object(Fig. 12-b,
c,
d),
so
the shapes are incorrect perspectively. Therefore
it
is
corrected using
a
view point with vanishing points
automatically.
Figure
12.
Creating
a
basic
shape
5.
Modeling Functions
5.2.
Cutting
5.1.
Thinning
At
first,
a
designer draws freehand lines onto
a
tablet
to become
a
perspective view for
a
proposed rectangu-
A basic shape is operated either adding parts
or
cut-
ting
off
pieces from the basic shape.
A
cutting plane is
decided from three points which are intersected points
59
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
of additional cutting lines on
a
basic shape. The most
probable plane is calculated if four or more points are
obtained (Fig.
13).
Basic shape Rough sketches
1
Cutting Rough sketches
2
Cutting Delete Erase cutting line Another view
Figure
13.
Cutting a basic shape
5
3.
Rounding
This section briefly introduces
a
method of rounding
operation from the polyhedra generated by the Sketch
Interpreter System. Rounding operation
is
carried out
by another program. The surfaces are represented by
the four-sided BCzier patches of the second degree. The
corners of polyhedra are represented by the patches
connected in
C1
continuity and the edges of polyhedra
are rounded by the patches coenected in
G1
continuity.
We can deal with the various degrees of rounding(Fig.
14).
6.
Examples
of
Product
Design
Figure 15 shows drawing procedures for an exam-
ple.
(l)and(2)
A
basic shape is at first constructed from
freehand lines. The size
of
a
model is automatically
evaluated in
a
computer from the sketch.
A
designer
need not take care of its practical size.
(3)
Another parallelepipeds are added. Edges are par-
allel to those
of
the basic shape.
(4)and(5) Cutting procedures are carried out by using
additional lines.
(6)
Expanding of
a
basic shape is carried out.
(7)
Modification of the adding parallelepipeds is carried
(8)
Models themselves can be rotated in the screen.
out.
7.
Conclusion
Figure
14.
Rounding Example
rithm of interactive line drawing, and reconstruction
of geometric models from sketches without scanning
sketches.
A
designer makes
3D
models like sculptur-
ing, to draw cutting lines on
3D
models directly.
Freehand Sketch Interpreter System is
a
new type
of
CAD
system which increases flexibility of data in-
put for
3D
models and will become the future trend
of
CAD
system. The system helps
a
designer on de-
signing industrial parts, whose shape is composed of
straight edges. Freehand lines are directly input to
a
computer through
a
tablet and instantly redrawn on
a
CRT screen with geometrically correct sketches. The
procedures work
so
intelligent that
a
designer feels free
of worrying about dimensioning of
a
spatial shape. The
designer can work at his workstation as the same man-
ner
as
drawing sketches with straight
lines
on
pqper.
Furthermore Sketch Interpreter transfers to another
program which can generate curved surfaces for pro-
viding high quality images of
3D
objects.
This paper introduces
a
user-friendly interactive
graphic system ”Sketch Interpreter”, thinning algo-
60
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
1.
Draw freehand lines
3.
Add another palallelepiped
2.
A
Basic Shape
4.
Modify a Basic Shape
5.
Many times modified drawing
6.
Expand palallelepipeds
7.
Complete a drawing
8.
Change view point
Figure
15.
Example
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
seaaces
[I]
Ullman D.G., Wood
S.,
Craig D.
:
The Impor-
tance
of
Drawing in the Mechanical Design Process,
Comp. and
Graph.14(2),1990,263-274.
[2]
Binh Pham
:
From CAD to Automation of Cre-
ative Design Process, CAGD94,Penang,Malaysia,
July. 1994.
[3]
B.
J. Hale,
R.
P.
Burton,
D.
R.
Olsen, and
W.
D.
Stout: A three-dimensional sketching environment
using two-dimensional perspective input, Journal
of
Imaging Science And Technology, pp.188-196,
March 1992.
[4]
K.
Kondo, F. Kimura, and
T.
Tajima: Estimation
of
a
Point of View with Perspective Drawing and
the Application, Journal of Information Processing
Society
of
Japan,Vol.29,No.7,July 1988.
[5]
K.
Kondo: Representation
of
3-D shapes with the
Interactive Rendering System,
a
doctor’s thesis,
University of TOKYO Japan, October 1987.
.
Akeo
et al.:
Computer-Assisted Product Design,
NICOGRAPM’93, pp.55-65, 1993.
[7]
Z
Chen and D-B Perng: Automatic reconstruction
of 3D solid objects from 2D orthographic views,
Pattern Recognition, pp.439-449, 1988.
[8]
T.
Fujiu
et aE.:
Usage
of
Heuristic Knowledge
for Solid Construction from Engineering Draft in
Orthographic View, NICOGRAPH’86, pp.49-58,
1986.
[9]
S.
Sugishita,
K.
Kondo, H.Sato, S.Shimada
:Sketch inteqweter system for geometric modelling,
CAGD’94, Penang, Malaysia,July. 1994.
[lo]
H.
Chiyokura: Solid Modelling (Japanese), Kogyo
Chosakai Publishing Co.,pp.l-31,1985.
[Ill
N.
Toriya: Invitation to 3D modelling (Japanese),
Myoritsu Syuppan Co., July 1991.
[
121
Masuda, Yamaguchi: Automatic Reconstruction
of
Polyhedron from Orthographic Views, Proc. of
Graphics and CAD Symposium, IPSJ, pp.105-114,
1992.
[13]
M.
Ono: Performance modelling
for
pen input,
Information Processing Society of Japan’93-HI-47-
L.
[14]
S.
Furushima
et
al.:
Generation
of
3-Dimensional
geometric model from rough-sketch,
NICOGRAPW’SO, pp.11-20, 1990.
[15] Robert
C.
Zeleznik
et
al.
:
SKETCH: An Interface
for Sketching 3D Scenes, SIGGRAPH ’96 Proceed-
ings, pp.163-170, 1996
[16]
S.
Sugishita, K. Kondo
et al.:
Interactive Free-
hand Sketch Interpreter for Geometric Modelling,
Symbiosis of Human and Artifact, pp.561-566, 1995
[17] K. Matsuda
et al.
:
User Interface for Sketch
In-
terpreter, Proceedings
of
the Eleventh Symposium
on Human Interface, pp.69-76, 1995
[IS]
XU
Zheng,
et al.
:
Free-form Surfaces Generation
in Freehand Sketch Interpreter, CADEM’95, Xi’an,
China, pp.261-266, 1995
62
Authorized licensed use limited to: Communication University of China. Downloaded on December 10, 2009 at 07:09 from IEEE Xplore. Restrictions apply.
... Three-dimensional (3D) reconstruction and line drawing interpretation is two-related area. There are many systems for reconstruction from single 2D line drawing (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20). They share common objective namely to guess or optimize the depth of each corner before reconstruct. ...
... Discussion on 3D reconstruction of a single line drawing can be categorized into three issues namely work on optimization-based (3),(5),(10), (13), (15), (16) and (17), work based on line labeling and constraint of boundary representation (1),(2), (8), (14), (18), (19) and development of a computer system that provide interface for on-line sketching and automatically reconstruct a 3D object (6),(7), (9), (11) and (12). ...
... The interpretation system uses graph-based constraint solver to establish the geometric relationship. Matsuda (12) developed a system called Sketch Interpreter that accepts sketches drawn on a CRT screen through a stylus pencil and a tablet. The system can correct geometric models in a computer even though input data of 2D sketches are drawn freehanded. ...
Article
Full-text available
This paper proposes a new framework to interpret a two-dimensional line drawing that represents a three-dimensional object. The framework synthesizes three algorithms namely thinning algorithms, chain code algorithm based on Freeman Chain Code (FCC), and a new corner detection algorithms that accept chain code series as its input. Linear system of two image regularities namely spatial structure and skewed symmetry is identified and created. Total least square method namely Bisection and Secant method has been used to solve the linear system. The solutions obtained are the depth-value of each visible junction. Matlab functions are created to represent and solve the linear system, and to visualize the three-dimensional object derived. Three stages identified in the new framework are pre-processing, two-dimensional feature extraction and deriving depth values of each junction. Experimental result shows that the new framework sucessfully interpret a cube and produce three-dimensional data. The comparison to the previous works on three-dimensional reconstruction is also discussed.
... By knowing their starting points and endpoints, all strokes can be identified. As a result, online sketch is more often used with other applications such as 2D model identification Sezgin et al., 2004;Ku et al., 2006) and 3D model reconstruction Shpitalni, 1995& 1996;Matsuda, 1997;Oh and Kim, 2003;Company et al., 2004;Masry et al., 2005). ...
... The process for interpreting line is similar for both but thinning process is required additionally for unordered points before identifying stroke. Curvature analysis (Jenskins and Martin, 1992), conic section equation , angular and distance test (Matsuda et al., 1997) and linearity test (Ku et al., 2006) are applied to determine type of lines whether they are straight lines, elliptic arcs, parabolic arcs or corners. ...
Article
Full-text available
Freehand sketch is a quick rough drawing for portraying ideas. Pencil and paper are inexpensive effective tools, commonly used during a conversation to elaborate an explanation. Overlapping lines provide additional information to the rough sketch and make it clearer. To make use of its information in downstream processes (e.g., input for generating 2D tool paths or for constructing 3D model), the overtraced sketch should be simplified first to a single line drawing. Presented in this paper is an approach for identifying a single line drawing from a paper-based overtraced freehand sketch. Key activities in this approach are thick line creation, corner detection, and line identification. After obtaining the overtraced image, proposing sliding window is used to form thick line drawing. In case of obtained non-uniform thick line, corner detection is required before forming a single line drawing. On the other hand, a single line drawing can be created directly from a uniform thick line drawing. Examples for demonstrating this approach are also presented.
... Freehand sketch interpreter system is a new type of CAD system which increases flexibility of data input for 3D models and will become the future trend of CAD system [1][2][3]. ...
Article
Full-text available
Freehand sketch interpreter system is a new type of CAD system which increases flexibility of data input for 3D models and will become the future trend of CAD system[1-3]. This paper discusses a freehand sketch interpreter system for constructing complex 3D solid models. The essential original idea of this paper is to create complex 3D object with a template topology library step by step from freehand strokes. The library is called T-LIB. The input strokes are analyzed as a 2D edge graph. Once the edge graph can be matched to a template of T-LIB, the 3D shape can be reconstructed (referring to Fig.1(a,b)) or modified (referring to Fig.1(c,d)). Designers can use or feel comfortable to use our system for realizing their idea sketches.
Chapter
CAD software and similar modeling systems can show 3D object models created by designers, but these advanced 3D systems demand precise data about the 3D object, especially in the steps of inputting data and making free-form surfaces. Although it has become common to use CAD/CAM systems to increase the efficiency of the industrial design process, traditional sketching is often more efficient in the early stages of concept design. Since it is difficult to design a CAD system which can create 3D objects from a concept sketch directly due to ambiguous information, recently sketch systems have been introduced to bridge the gap between concept design and computer-based modeling programs, combining some of the features of a pencil-and-paper sketch. It has now become a big area of research of increasing the flexibility of 3D data input techniques for CAD systems. Concept Sketch is a concept design process for the designer. It is used to draw contours of products on paper by pencil step by step. In order to get 3D shapes, it is required to recognize these shapes from the concept sketch. Designers can not use or feel very uncomfortable to use these systems for realizing their idea sketches. This survey paper deals with a freehand sketch based geometric modeling for constructing complex 3D objects. In this talk, many useful sketch systems and methods are classified. The following sketch systems are shown: (1) A sketch-based modeling system based on descriptive geometry is proposed to make a model of a variety of 3D solid objects and surface objects. The main goal of sketch-based systems is to allow the creation of 3D models by using strokes extracted from user input and/or existing drawing scans. (2) The contour line method to propose a sketch interpreter system for designing 3D freeform objects. The essential idea is the use of freeform stroke as an expressive design tool. The Freeform stroke is classified into three types: outline (contour line), shading line and cross section line. The outline with shading lines are used to generate the basic rounded 3D shapes. The shading lines and cross section lines are used as control lines which are drawn inside the contour for modifying the basic 3D shape. (3) A method using a template topology library as an essential tool to reconstruct and modify 3D objects with sketch lines. 3D objects can be generated in a basic 3D shape reconstruction procedure, or modified in a 3D shape modification procedure. (4) A Sketch modeling of Implicit Surfaces by using cutting method and three orthographic views. We developed a cutting method using implicit surfaces based on a Cube. Characteristic point is blending function compare with our three proposed sketches interpreter system. (5) Digital sculpting with history management of strokes. This is a new digital sculpting system based on the history of stroke input. In these systems we are interested in interpretation, where the computer creates the 3D objects step by step. The designer can feel more comfortable to quickly evaluate their ideas for designing a new object compared with the traditional CAD system.
Article
The complicated menu-based, command line, or scriptable interface of most simulation tools creates an unnatural working environment for model development and analysis. Even engineers and scientists familiar with such systems spend significant time converting hand-drawn designs into digital models for simulation. They then spend more time preparing these results for visualization. These steps lead to many hours exhausted on transcribing, rather than interpreting results. We present a tool that allows users to sketch mechanical systems using standard engineering representation, to simulate the responses of those systems, and to view the results of those simulations, all on the same "sketch-pad.".
Article
CAD software and similar modeling systems can show 3D object models created by designers, but these advanced 3D systems demand precise data about the 3D object, especially in the steps of inputting data and making free-form surfaces. Although it has become common to use CAD/CAM systems to increase the efficiency of the industrial design process, traditional sketching is often more efficient in the early stages of concept design. Since it is difficult to design a CAD system which can create 3D objects from a concept sketch directly due to ambiguous information, recently sketch systems have been introduced to bridge the gap between concept design and computer-based modeling programs, combining some of the features of a pencil-and-paper sketch. This survey paper deals with a freehand sketch based geometric modeling for constructing complex 3D objects. In addition, many useful sketch systems and methods are classified. The following sketch systems are shown: (1) A sketch-based modeling system based on descriptive geometry is proposed to generate a model from a variety of 3D solid objects and surface objects. (2) The contour line method to propose a sketch interpreter system for designing 3D freeform objects. (3) A method using a template topology library as an essential tool to reconstruct and modify 3D objects with sketch lines.
Article
Freehand sketch is a natural and intuitive communication channel for idea expression. Lines are drawn one after another to create the outline before additional lines are often drawn over the existing ones to make a sketch clearer. Typically, the sketch is transformed to be a 3D CAD model by a designer for use in subsequent operations. Sketch-based modeling has been researched to support this transformation but mainly for an online sketch on a digital device. For an offline sketch on a paper, commonly found used in practice, sketch-based modeling remains a challenge because a scanned image conceals enriched information on the sketch in a batch of data points. This paper proposes an approach for reconstructing a 3D model from a paper-based overtraced freehand sketch. Two main modules in this approach are single-line drawing identification and 3D reconstruction. The first module where image processing technique is applied is for transforming a paper-based overtraced freehand sketch to be a single-line drawing in order to generate more information about the sketch (i.e., the number of lines and their starting points and endpoints). The second module where progressive reconstruction approach with cubic corner is applied is for reconstructing a 3D model from the obtained single-line drawing. Steps to be taken in both modules have been formalized. The approach has been successfully implemented on LabVIEW, and tested with several samples.
Article
This paper describes a new method for recognizing overtraced strokes to 2D geometric primitives, which are further interpreted as 2D line drawings. This method can support rapid grouping and fitting of overtraced polylines or conic curves based on the classified characteristics of each stroke during its preprocessing stage. The orientation and its endpoints of a classified stroke are used in the stroke grouping process. The grouped strokes are then fitted with 2D geometry. This method can deal with overtraced sketch strokes in both solid and dash linestyles, fit grouped polylines as a whole polyline and simply fit conic strokes without computing the direction of a stroke. It avoids losing joint information due to segmentation of a polyline into line-segments. The proposed method has been tested with our freehand sketch recognition system (FSR), which is robust and easier to use by removing some limitations embedded with most existing sketching systems which only accept non-overtraced stroke drawing. The test results showed that the proposed method can support freehand sketching based conceptual design with no limitations on drawing sequence, directions and overtraced cases while achieving a satisfactory interpretation rate.
Conference Paper
Full-text available
The evolution of Computer Aided Design (CAD) and its application to solve engineering problem such as interpreting engineering sketch is started with the invention of Sketch Interpreter in 1963 by Sutherland. Since then, previous research on this topics have presented several frameworks to read, analyze, represent, validate, reconstruct, and visualize an object from engineering sketch to a valid solid object. This paper presents a new framework in a complete cycle of transformation object from engineering sketch to valid solid objects. It consists of four phases: the interpreting of engineering sketch, representing the regular line drawing, labeling and validating the line drawing, and reconstructing the validated line drawing. The framework has been tested to few objects and has been implemented by developing a prototype system. The system starts from reading two-dimensional (2D) engineering sketch, thinning the sketch, extracting the feature and represents it using the concept of chain code-scheme. The generated regular line drawing is then labeled using Huffman-Clowes line labeling to identify its validity in representing three-dimensional (3D) object and rejects the impossible drawing. Neural network with back propagation is employed for the subsequently step in deriving the z-values of the visible junctions and the hidden junction of the object. Finally, the reconstructed object is represented as mathematical modeling and visualized. As a conclusion, the combination of thinning algorithm, chain-code scheme, line labeling, neural network, and mathematical modeling in the proposed framework has generated a new invention in the development of sketch interpreter.
Article
Full-text available
This paper is a study on the importance of drawing (both formal drafting and informal sketching) during the process of mechanical design. Five hypotheses, focused on the types of drawings, their necessity in mechanical problem solving, and their relation to the external representation medium, are presented and supported. Support is through referenced studies in other domains and the results of protocol studies performed on five mechanical designers. Videotapes of all the marks-on-paper made by designers in representative sections of the design process were studied in detail for their type and purpose. The resulting data is supportive of the hypotheses. These results also give requirements for future computer aided design tools and graphics education, and goals for further studies.
Conference Paper
Sketching communicates ideas rapidly through approximate visual images with low overhead (pencil and paper), no need for precision or specialized knowledge, and ease of low-level correction and revision. In contrast, most 3D computer modeling systems are good at generating arbitrary views of precise 3D models and support high-level editing and revision. The SKETCH application described in this paper attempts to combine the advantages of each in order to create an environment for rapidly conceptualizing and editing approximate 3D scenes. To achieve this, SKETCH uses simple non-photorealistic rendering and a purely gestural interface based on simplified line drawings of primitives that allows all operations to be specified within the 3D world.
Article
This paper describes a method of reconstructing a polyhedron from orthographic views. First, a wire frame model is created. This is achieved by referring to the corresponding relationship among three orthographic views. Then a surface model is generated by making loops out of coplanar edges. Such a surface model consists of ghost figures. This paper deals with the ghost figure eliminating procedure as determination of flags, flags referring to nodes in the constraint relationship network. The flag of each node in this network represent true, false or undecided. With this approach, many rules in consistency checking and local value propagation are systematically organized, and the consistency managing procedure is completely isolated from the search procedure. As a result, the systematic rules can be adopted at any time in the search process when needed. The method implemented in this paper shows the way to use consistency checking and local value propagation effectively.
Article
This paper introduces a user-friendly man-machine interface to create geometric models by inputting sketches on a CRT screen. The screen provides coordinate axes of a 3D space drawn perspectively with vanishing points which are familiar to most designers as sketch aids. The system, Sketch Interpreter, is constructed in terms of three characteristic operations on a workstation with a pointing device: (1) to recognize handwritten sketches on the screen: freehand lines are drawn by pen operations from which a computer can construct a geometric model. This first operation computes mathematical parameters for the projective transformation; (2) to construct additional geometric models by inputting more sketches drawn perspectively; and (3) to redraw modified geometric models replacing the sketched lines. The interactive methods are suited to realistically constructing any geometric shape that a designer imagines.
Article
This paper presents a new method to deal with idea sketches for inputting geometric models at a workstation. The idea sketches are drawn on a CRT screen with a stulus pen and a tablet by designers at an initial stage of design procedures. They can keep their drawing styles at a workstation as the same manner as using a pen on paper. The system 'Sketch Interpreter1 can create correct geometric models in a computer even though input idea sketches are incorrect perspectively. Data created are transferred to an advanced 3D-CAD system. The system is applied as a front-end processer for a design practice.
Article
An algorithm which can automatically construct 3D solid objects from 2D orthographic views is proposed. The views contain geometric information of lines, circles or circular arcs. The reconstructed objects may be polyhedra, cylinders, partial cylinders and their composites. The reconstruction process consists of three phases: decomposition, reconstruction and composition. First, the inputted drawing is decomposed into several predefined types of subviews, then a translation sweep operation, perhaps followed by a plane-cutting operation, is used to reconstruct the corresponding subpart for each set of subviews. Finally, the volume enclosure relationships between these subparts are utilized to compose the final part.
Article
Sketching communicates ideas rapidly through approximate visual images with low overhead (pencil and paper), no need for precision or specialized knowledge, and ease of low-level correction and revision. In contrast, most 3D computer modeling systems are good at generating arbitrary views of precise 3D models and support high-level editing and revision. TheSKETCH application described in this paper attempts to combine the advantages of each in order to create an environment for rapidly conceptualizing and editing approximate 3D scenes. To achieve this, SKETCH uses simple non-photorealistic rendering and a purely gestural interface based on simplified line drawings of primitives that allows all operations to be specified within the 3D world.