Towards a Natural Interaction for Rapid Prototyping of Parametric

Closets

an Rodr

ıguez

, Celso Campos

, Enrique Barreiro

, Jorge V

azquez

and Roc

ıo Veiga

Department of Computer Science, University of Vigo, Campus As Lagoas 32004, Ourense, Spain

Redegal S.L., Avda. de la Habana, 43 1

Izda. 32004, Ourense, Spain

Keywords:

Procedural Modelling, CAD, Multi-touch Devices, Natural Interaction, Rapid Prototyping, Closets.

Abstract:

Custom closets customers usually lack the expertise to design their desired closet. The use of a software tool

that incorporates the design expertise might allow them to sketch their ideas without third party oversight.

While architecture related applications are very powerful and make it possible to have an accurate graphical

representation of real world, these Computer-Aided Design (CAD) tools demand high technical training and

their use in the engineering ﬁeld is generic. Our work, framed within the development of a rapid prototyping

tool of custom closets called Sketch Arm, introduces an interaction proposal that seeks to minimize the time

and complexity associated with typical operations in CAD. Supported by multi-touch capabilities of today’s

mobile devices, our system allows users to manipulate the closet structure directly in real time by using a

minimalist graphical user interface (GUI) together with a touch interface based on easily recognizable gesture

patterns. Direct on-screen data manipulation is very attractive to the end-user in general and highly beneﬁcial

for novice IT users in particular while gesture-based interaction is extremely intuitive for performing artistic

tasks. Throughout the article, we will describe in depth the designed interaction system and we will present

the original solutions we made to common problems in touch systems like accuracy and occlusion.

1 INTRODUCTION

Custom closets manufacturing arises in response to

the spatial limitations that nowadays many of the ex-

isting housing present, improving the organization

and use of available space in living quarters.

The customer usually lacks the expertise to deal

with the design process independently and is assisted

by the manufacturer for the creation of an initial pro-

totype together. This model is time consuming and

involves a signiﬁcant increase in the ﬁnal manufactur-

ing cost. The use of a software tool that incorporates

the design expertise, will allow the user to sketch his

ideas without third party oversight.

To that end, there are many computer applications

that facilitate the design process, enabling us to obtain

a graphical representation of the real world. These

Computer-Aided Design (CAD) tools, although very

powerful, demand high technical training and their

use in the engineering ﬁeld is generic.

To address those issues, we have developed a rapid

prototyping application for designing custom closets,

named Sketch Arm. The main purpose of the system

is to facilitate the operations involved in the creative

process. We propose a direct interaction system that,

supported by multi-touch capabilities of current mo-

bile devices, allows the user to manipulate the closet

structure in real time, combining the use of a com-

mon graphical user interface (GUI) with the use of a

gesture-based natural interface.

Touching and manipulating data directly on the

screen, without using any additional input device, is

very attractive to the end-user: this direct manipu-

lation style mainly beneﬁts novice IT users (Albins-

son and Zhai, 2003; Benko et al., 2006) thanks to a

gentle learning curve. For its part, gestured-based in-

teraction is an extremely intuitive interface —as it is

based on inherent human capabilities— for handling

applications that involve any artistic work. However,

the mere incorporation of both interaction methods to

human-machine communication process does not en-

sure end-user satisfaction. We must design a global

interaction in which each element has a purpose and

brings added value to user experience.

According to the established goal, we organize

the rest of the paper as follows. The second sec-

tion presents a brief summary of several supporting

works and techniques. Then, we provide an overview

409

Rodríguez I., Campos C., Barreiro E., Vázquez J. and Veiga R..

Towards a Natural Interaction for Rapid Prototyping of Parametric Closets.

DOI: 10.5220/0004633104090416

In Proceedings of the 9th International Conference on Computer Graphics Theory and Applications (GRAPP-2014), pages 409-416

ISBN: 978-989-758-002-4

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

of Sketch Arm in order to put our work in context.

The remaining two sections are focused on the user-

system interaction: the fourth one provides a brief

description of the different stages in the interactive

closet design process, and the next one describes in

more detail the various elements of the interaction

system proposed to support design tasks. Finally, we

present the most signiﬁcant conclusions of our work

and we identify future lines of action.

2 BACKGROUND

Multi-touch technology has been extensive studied in

the past 20 years, in which many authors have focused

their efforts, among other things, on (i) reducing the

manufacturing cost of the supporting hardware (Han,

2005), (ii) enhancing its portability (Pinhanez et al.,

2003) and sensitivity (Dietz and Leigh, 2001; Reki-

moto, 2002), and (iii) developing software systems

based on this new interface.

With regard to the last point, there are already sev-

eral works in the literature that, instead of creating

these systems as mere testbeds for the design of a new

interaction proposal, seek to develop really useful

tools. As an example, we can mention applications for

purposes as diverse as medical imaging (Lundstrom

et al., 2011), sketch design (Bae et al., 2008) and

animation (Quevedo-Fern

andez and Martens, 2012),

or information visualization in museums (Hornecker,

2008).

However, almost all of these applications are

aimed to casual users —users with low IT skills—

and its exploitation in professional contexts is usually

neglected. In this sense, it would be interesting to

examine to what extent this interaction may allow us

to simplify the use of professional tools —eminently

complex—, either to enable expert users to carry out

their work expeditiously or to bring its use closer to

novice users.

To that end, our paper presents a novel and attrac-

tive interaction proposal which minimizes the time

and complexity normally associated with typical op-

erations in CAD environments, within the profes-

sional context of furniture industry. While direct-

touch interaction provides a greater interactive rich-

ness than traditional WIMP (Windows, Icons, Menus,

Pointer) systems, the implementation of this touch-

based interaction for common CAD tasks is not triv-

ial. Among these tasks, the selection operation —

highly performed in this kind of tools— is especially

problematic. Speciﬁcally, the two main problems tra-

ditionally associated with using the ﬁnger as a point-

ing and selection tool are accuracy and occlusion.

Multiple authors have sought to determine the

ideal size of GUI widgets in order to minimize the

accuracy problem. Thus, Hall et al. point out in their

1988 study (Hall et al., 1988) that this accuracy can

be 99.2% for targets with a size of 26 mm. Subse-

quently, however, A. Sears and B. Schneiderman note

in his work (Sears and Shneiderman, 1991) that ac-

curacy is maximized when the target has a size of 32

pixels —13.8 mm x 17.9 mm—. This study still holds

true, being borned by different recent papers (Dan-

dekar et al., 2003; Wang and Ren, 2009).

There are several additional approaches in the lit-

erature that have also attempted to solve the accuracy

problem in traditional touch systems by (i) using a

wider selection area or hot spot instead of the mouse

cursor (Kabbash and Buxton, 1995), (ii) using both

selection methods (Worden et al., 1997), or (iii) using

resizable widgets (Bederson, 2000), that can dynam-

ically change their size in order to provide a larger

interaction area to users.

For their part, in relation to the occlusion problem,

some authors have primarily focused their efforts on

preventing the user from hiding the on-screen cursor

with his ﬁngers. The methods proposed to achieve

this, range from applying a small offset above the

user’s ﬁngertip —ﬁxed in Take Off, a technique in-

troduced by Potter et al. in (Potter et al., 1988), and

dynamic in the case of the technique named Dual Fin-

ger Offset, introduced by Benko et al. in (Benko et al.,

2006)— to the cursor, to the use of an on-screen call-

out for displaying a copy of the occluded screen area

(Vogel and Baudisch, 2007).

3 SKETCH ARM. SYSTEM

OVERVIEW

Sketch Arm is organized into three distinct but closely

related subsystems: (A) interactive sketching subsys-

tem, (B) automatic design generation subsystem and

two subsystems are responsible for generating the 3D

geometric information that represents the closet. The

third one is responsible for providing a visual repre-

sentation of the closet, created from that information,

as the ﬁnal output of the entire system. Figure 1 shows

a diagram of the overall architecture of Sketch Arm.

As shown in Figure 1, subsystems A and B receive

information from existing knowledge bases and data

obtained from user interaction as input, and they both

provide as output a geometric model that character-

izes the closet structure. However, both subsystems

deal with the creation of the model by signiﬁcantly

different strategies:

GRAPP2014-InternationalConferenceonComputerGraphicsTheoryandApplications

410

!""#$$%&' #$

!()'* +#&!"+', #)$- #+" .'* /)$01$2$+# 3

&445

6789:;

!(<(

=44>)6789:;

!(?(

':@6>:AB)

BAC4D@

6789:;

!(E(

!0+% 3!+'")

/#* #&!+' %*)

$01 $2$+ #3

=#$'/*)

"%* $+& !'* +$

$'F #)

"%* $+& !'* +$

?=)&#*= #&' */)

$01 $2$+ #3

0$# &G+%0".)$0& H!"#)

'*+ #&! "+'%*)=!+!

"I%$#+)

='3 #*$ '%*$

!3% 0*+ )%H)

"I%+ .#$)+%)1#)

$+%& #=

?=);6456@>8J)8:K4>5A94:)4K)@L6)JB4M6@

"45NA>@56:@M)

5A@JL8:;

':7DM@>C)M@A:7A>7)M8O6

/6456@>8J)

8:K4>5A94:)4K)@L6)

AJJ6MM4>86M

$@A:7A>7)M8O6)4K)AJJ6MM4>86M

"45NA>@56:@GAJJ6MM4>C)5A@JL8:;

Figure 1: Overall architecture of Sketch Arm.

• The interactive editing subsystem takes as input

data collected from the interaction between the

user and the multi-touch surface of the supporting

hardware device. This subsystem is intended to

allow the end-user to emulate the creative process

undertaken by professional designers and it pro-

vides him with the necessary tools to create and

customize in real time the various constituent ele-

ments of the closet.

• The automatic generation subsystem demands al-

most no user intervention: the individual must

provide only the dimensions of the closet and the

amount of clothing he wants to store in percent-

age terms. This subsystem is able to generate

autonomously different design proposals to meet

both users’ storage needs and, as far as possible,

the design and dimensional constraints contained

in the system. The paper (Rodr

ıguez et al., 2013)

provides a detailed description of this subsystem

and the conﬁguration of the proposed stochastic

optimization algorithm for the automatic layout

generation.

In the rest of the paper we will focus on the in-

teractive editing subsystem. We will present the dif-

ferent tasks involved in the prototyping process and

describe in more detail the interaction system we de-

signed to support them.

4 DESIGN FLOW IN THE

INTERACTIVE EDITING

PROCESS

The design process consists of a set of operations for

the three-dimensional parametric modeling of clos-

ets. This process is broken down into three sequential

stages and each stage results in a functional module

that gathers the relevant operations. Thus, we have a

room editing module, an internal layout editing mod-

ule and a door editing module.

4.1 Room Editing Module

The primary aim of this module is to obtain a geomet-

ric model of the external structure of the closet from

a basic outline of the room. The user deﬁnes this out-

line by creating the walls and pillars that make up the

structural shape of the room. Then, the position, ori-

entation and size of the closet are directly calculated,

taking as a reference the geometry of the selected sup-

porting wall. To do this:

• The origin of the closet is translated to the position

of the origin point of the wall.

• The closet is rotated towards the direction of the

normal vector of the wall.

• The width and height of the closet are deter-

mined by the corresponding dimensions of the

wall. Depth, however, is constant and takes the

standard value in closet manufacturing industry.

4.2 Internal Layout Editing Module

The design is completed by incorporating additional

geometric elements to the basic model created in the

previous stage. These representative elements of the

internal structure of the closet include vertical and

horizontal dividing elements, and instantiable ele-

ments or accessories —bars, drawers, etc.—.

While the addition of an instantiable element is

just a simple assignment operation, introducing a di-

viding element is a complex process that causes the

segmentation of a compartment into two new child

compartments. This process involves the creation of

new structural elements and also the computation of

the transformation matrices needed to adjust each ele-

ment to the local system deﬁned by the interior of the

parent compartment. Transformations are calculated,

therefore, from the scale, position and orientation of

this compartment.

4.3 Door Editing Module

Once the structure of the closet carcass has been de-

tailed, all that remains is to create the front doors.

This process involves three stages:

1. Constructing an initial basic geometric model for

each door.

2. Dividing each door into panels.

3. Conﬁguring the visual appearance of each panel.

TowardsaNaturalInteractionforRapidPrototypingofParametricClosets

411

The construction of the initial model of the door is

automatic and completely transparent to the end-user:

the system is responsible for deﬁning the geometry

of the model and computing the scale, translation and

rotation matrices for proper rendering. However, the

division stage demands user intervention like the edi-

tion of the internal layout of the closet, but it only al-

lows horizontal dividers. Finally, in the visual appear-

ance conﬁguration stage the user can assign a material

or color to each created panel as a texture of its front

face.

5 INTERACTION SYSTEM

Sketch Arm interaction system rests on three major

guidelines or design principles:

1. A simpler GUI

2. A touch-based interface for the direct manipula-

tion of the elements in the design.

3. The minimization and simpliﬁcation of mode

switching.

Throughout this section we will present in greater

detail the work we have carried out in relation to the

outlined design principles and the solutions we pro-

pose to different problems associated with using the

ﬁngers for pointing and selecting.

5.1 Simplifying the GUI

The design of the graphical interface aims to give the

user an access as immediate as possible to system

functionality. We remove the superﬂuous elements of

the application to get a clean, clear and minimalist in-

terface that cedes part of its prominence to provide the

end-user with a seamless and superior experience.

The tool is designed as a single landscape view

that includes all the graphical elements of the user in-

terface (Figure 2).

Drawing canvas

Navigation

bar

Toolbar

Accessories

Visual ﬁnishes for

door panels

Figure 2: Layout of Sketch Arm GUI.

These elements are displayed on screen according

to the usual layout of iOS Apps:

• The toolbar gives the user access to functionality

for creating and modifying the different compo-

nents of the closet.

• The drawing canvas is where the user performs

the tasks involved in the design process. Through

this central panel, the system gives constant feed-

back to the user, displaying in real time the design

resulted from each operation.

• On the right of the drawing canvas we include a

panel that provides direct access to both visual ﬁn-

ishes for the different panels of the front doors and

accessories that can be included inside the com-

partments of the carcass.

• The navigation bar allows the user to toggle be-

tween the different views associated with each

functional module.

The design of the graphical elements must be aes-

thetically pleasing in order to create an engaging ex-

perience for the end-user. In touch systems it is

also necessary to focus efforts on appropriately sizing

these elements so as not to hinder the interaction. On

the basis of the works (Dandekar et al., 2003; Sears

and Shneiderman, 1991; Wang and Ren, 2009) re-

ferred in the state of the art, we adequate the GUI to

ﬁt the touch interaction style by following the seen

dimensional guidelines. Thus, we set the toolbar but-

tons size to 40 pixels per side —14.1 x 14.1 mm—,

the right panel buttons size to 54 pixels per side —

19.1 x 19.1 mm—, and the width of the buttons on

the navigation bar to 76 pixels and its height to 44

pixels —26.8 x 15.5 mm—.

5.2 Touch-based Interface for the Direct

Manipulation of the Design

Elements

The interaction between the user and the application

is mainly channeled through the central drawing

canvas. We assume that the ﬁrst ﬁnger that touches

the canvas, called primary ﬁnger, is the index ﬁnger

of the dominant hand. Any additional touch is

assigned to any other ﬁnger of either hand. We deﬁne

four action templates to support the tasks involved in

the design process: (i) single or multiple taps on any

GUI control, (ii) single tap on any design element

and then single tap on any GUI control, (iii) single

tap on any GUI control and then single tap on any

design element, and (iv) single tap on any design

element and drag.

i. Single or multiple taps on any GUI control. The

user taps with only one ﬁnger on a GUI control to

GRAPP2014-InternationalConferenceonComputerGraphicsTheoryandApplications

412

activate the corresponding tool and thus trigger the

actions associated with it. This template is reserved

for the use of certain functionalities —primarily to

create new objects in the scene— that are imple-

mented as automatisms and require no additional

user interaction to the simple activation of the tool.

As seen in Figure 3, this action template is visible,

for example, when the user adds a new door to the

design: when the button is single tapped by an user,

the system performs the construction of the geometric

model of the new door autonomously.

Figure 3: Door creation process. Each single tap on the

associated button triggers the calculations needed to para-

metrically deﬁne a new door and change the position and

scale parameters of previously created doors.

ii. Single tap on any design element and then

single tap on any GUI control. Firstly, the user

single taps an object to select it. This way, the

individual intentionally points to the object on which

he wants to perform an operation. Then, the subject

uses his ﬁnger again to tap the button associated with

that operation to modify the selected object or as

reference for adding a new element to the design.

Thus, for example, to include a horizontal division

inside the closet, the user has to ﬁrst select a target

compartment and then tap the shelves creation button.

As shown in Figure 4, this interaction template is

also evident in most of the editing operations, like the

deletion of an element.

Figure 4: Deletion of an object from the design.

iii. Single tap on any GUI control and then single

tap on any design element. There are some tools

that, once selected, require an extra touch on the

drawing canvas to create a new object at a speciﬁc

location in the scene. The user explicitly provides

the location by pointing his ﬁnger to a certain spot,

which is an eminently natural gesture. The system

translates the 2D coordinates of the touch event to

the position coordinates of the target location in the

3D Cartesian space. This template is visible, among

other tasks, in wall creation: the user ﬁrst selects the

tool (Figure 5.1) and then creates the endpoints of the

walls with successive single taps (Figure 5.2).

Figure 5: Creation of the walls that make up the shape of

the room.

iv. Single tap on any design element and drag. Be-

yond speciﬁc object creation and deletion tools, there

are some other ones that must be included in any CAD

or rapid prototyping application. We are talking about

translation and scaling tools. These operations, as we

will see next, are closely related in our system.

To perform the translation of an object, the user

must ﬁrst tap the object to select it (Figure 6.1) and

then, without lifting his ﬁnger from the touch surface,

drag it across the screen to the target position (Fig-

ure 6.2). The translation ends when the user lifts the

ﬁnger from the touchscreen (Figure 6.3).

As the user drags his ﬁnger across the drawing

canvas, the object slides in such a way that it re-

mains under the ﬁngertip. This direct manipulation

provides the user with an engaging experience, since

the movement of the selected object thus conceived is

predictable and transmits to the user the sensation of

actually moving the object.

!"#

!$#

!%#

Figure 6: Translation of an object across the drawing can-

vas.

The translation of the object does not exclusively

affect its position. We implemented an algorithm

called indirect scaling that modiﬁes the dimensions

of the sketch elements that have some connection or

relationship with the translated object. Thus, for ex-

ample, when the user moves a divisor, the system au-

tomatically resizes both adjacent compartments and

the accessories inside them.

Indirect scaling is signiﬁcantly different from the

approach commonly used in traditional CAD tools in

which user must explicitly enter the dimensions for

TowardsaNaturalInteractionforRapidPrototypingofParametricClosets

413

each object by using a physical keyboard. The main

alternative to physical keyboard in touch interfaces is

the use of an on-screen keyboard, widely reviewed

in the literature (Zhai et al., 2000) and extensively

implemented in recent multi-touch mobile devices.

Nevertheless, we reject the use of the touch keyboard

due to its ergonomic and feedback problems (Lewis

et al., 1997) that adversely affect the end-user’s per-

formance and comfort.

Our scaling mechanism enables an efﬁcient ob-

ject sizing in touch-based systems but it is inherently

less accurate than directly typing the different values

with a physical keyboard. To improve this, we im-

plemented a non-intrusive accuracy mechanism. The

user can naturally activate the accuracy mode while he

is sliding an object across the drawing canvas. To do

this, the user must keep the ﬁngertip still on the object

for 1000 ms. Thereafter, a smaller scale is applied to

any object translation, thus increasing the movement

accuracy. The user ends this mode by tapping any-

where in the canvas except on the object itself.

5.3 Scene Elements Selection

Mechanism

The introduced action templates largely rely on ﬁrst

selecting an element of the sketch. This is common to

many typical CAD tasks, which also require the selec-

tion of a speciﬁc object or entity to trigger a particular

process.

While in desktop CAD tools the individual uses

the mouse cursor, Sketch Arm’s touch nature and the

consequent absence of the on-screen pointer force the

individual to use his ﬁnger as a pointing and selecting

device, causing accuracy and occlusion problems. To

solve both problems, we have implemented a number

of speciﬁc mechanisms that seek to address them

from different angles.

Additional viewport for displaying relevant infor-

mation. Inspired by the Shift technique, introduced

by Vogel and Baudisch (Vogel and Baudisch, 2007),

our method also consists in including a small on-

screen additional viewport when one ﬁnger taps the

touch surface. This feature is only enabled for cer-

tain operations that demand high accuracy. However,

unlike Shift, the viewport does not display a copy of

the area occluded by the ﬁnger; it provides text in-

formation that is relevant in the operation context and

fundamental for its successful completion. Figure 7

shows an example of the use of this technique to sup-

port the creation of walls.

After creating the ﬁrst end of the wall, the system

provides a real-time preview of it as the user drags his

Figure 7: Using the additional viewport to enhance the wall

creation process.

ﬁnger on the drawing canvas to set the other end. As

shown in Figure 7, the ﬁnger itself prevents the indi-

vidual from seeing the exact position of the second

end, causing the loss of an essential reference for the

successful creation of the wall. In order to prevent

this, the additional viewport shows the exact length of

the wall with an accuracy of two decimal places.

The auxiliary display area is shown when one

ﬁnger taps anywhere in the drawing canvas and it

is located in a position near the contact point. The

magnitude of displacement between the ﬁnger and

the viewport positions is constant. Thanks to this

we avoid changing the user’s visual attention during

the object translation. For its part, the direction and

orientation of the offset vary depending on the ﬁnger

position in order to ensure that the viewport is always

shown within the ﬁxed limits of the canvas.

Non-visible expansion of the selectable area of an

object. This mechanism seeks to solve the accuracy

problem without visually change the size of the ob-

jects in the scene. A non-visible rectangular area is

associated to each element of the sketch and acts as

its selectable area: the graphical information of the

area is not displayed on screen, but it is rendered to an

off-screen buffer. Thus, we can increase the on-screen

target size in order to facilitate the selection task while

maintaining a realistic representation of the objects.

!"#

!$#

!%#

Figure 8: Expansion of the selectable area of the different

elements of the closet carcass. (a) shows a representation of

the color information for the corresponding selectable areas.

(b) shows the on-screen displaying of the closet. In (c) we

provide a graphic of how the selectable area expands the

area of a closet shelf.

As shown in Figure 8, there is a clear center align-

ment between each element of the sketch and its cor-

responding selectable area. To expand the selectable

area we assign to it dimensions that exceed the size of

GRAPP2014-InternationalConferenceonComputerGraphicsTheoryandApplications

414

the object itself.

Our selection algorithm is based on the color

picking method, common in OpenGL or DirectX.

To support this algorithm a unique color is assigned

to each selectable area (Figure 8.a). A unique ID,

consisting of the RGB color values of the area,

is associated to the corresponding object. Having

established this relationship, the algorithm is able to

extract the color information of the exact pixel where

the ﬁnger contacts with the selection area and then

search the object whose ID matches the RGB data.

Increasing information space by using zoom. The

zoom tool allows the user to adjust the information

space to an appropriate size for selecting on-screen

items comfortably. The implemented algorithm is

based on the use of the standard pinch multi-touch

gestures to make the zoom mode activation fully

transparent to the user. This mode is automatically

activated when the individual begins to drag two ﬁn-

gers describing a pinch-in or pinch-out gesture. The

amount of zoom is calculated in real time from the

distance between those two ﬁngers.

5.4 Minimizing and Facilitating Mode

Switching

Generally, mode switching means a change of con-

text that usually involves interrupting the workﬂow.

In this respect, it will be possible to streamline the

user’s tasks to the extent we are able to minimize the

need to change between modes.

However, the wide range of features and func-

tions that software design tools usually provide, sig-

niﬁcantly hinders this minimization: when the user

interacts with this kind of tools, it is usual that he in-

tentionally selects the corresponding dedicated con-

trol in order to activate the new mode.

Our approach is based on two strategies that seek

to simplify the alternation between them. These

strategies consist of (i) providing user transparent ac-

cess to frequently used tools and (ii) using multi-

touch gestures for switching between the different

modes associated with these tools.

Sketch Arm provides fully transparent access to

selection, translation and, as we have already seen,

zoom tools:

• The selection mode is automatically enabled

when the user ends any operation and disables its

corresponding control.

• When the selection mode is enabled and an ob-

ject is selected, a drag gesture with one ﬁnger

enables the translation mode, as previously indi-

cated. When the user lifts the ﬁnger from the

touchscreen, the translation mode ends and the se-

lection mode is automatically enabled again.

• Sketch Arm allows switching to zoom mode at any

moment, regardless of the current mode, when the

user performs a pinch gesture. As soon as the user

lifts one ﬁnger, the zoom mode ends and the ex-

isting context before its activation is recovered.

6 CONCLUSIONS AND FUTURE

RESEARCH

As a result of the work we have done, Sketch Arm

brings closet design to non-professional users by pro-

viding a touch-based experience and a virtually trans-

parent interface that allows them to interact with the

system in a natural manner. A user interface designed

in this way, facilitates learning through the design

process and generates on the individual the sensation

of mastering the tools after little training. Bellow we

summarize the main contributions of our work:

• Graphic user interface with an attractive and min-

imalist design.

• Design of control elements suitable for touch in-

teraction.

• Touch-based action templates for a natural inter-

action in sketching tasks.

• A new approach to object scaling in touch-based

systems called indirect scaling.

• The design and implementation of speciﬁc mech-

anisms to avoid occlusion and accuracy problems,

caused by the use of a ﬁnger as a pointing device.

Taking the Shift technique as a reference, we use

an additional viewport for displaying relevant in-

formation in certain operations; we introduce an

innovative technique that seeks to solve the ac-

curacy problem by associating to each on-screen

object a non-visible larger selectable area; and, ﬁ-

nally, we simplify the switch between modes by

using multi-touch gestures.

For further work, it will be necessary to conduct

empirical studies in order to evaluate the goodness

of the proposed interaction system in terms of the

performance and error rate obtained by professional

and novice users when performing certain tasks in

the design process. It will also be desirable to make

a comparative study to determine which kind of

system, touch or desktop, is the most suitable for

each considered user proﬁle based on the individuals’

performance and their subjective opinion.

TowardsaNaturalInteractionforRapidPrototypingofParametricClosets

415

ACKNOWLEDGEMENTS

This work was supported by project 10DPI305002PR

of Xunta de Galicia and by project ITC-20133062 co-

funded by the Ministry of Economy and Competitive-

ness through CDTI within the Innterconecta FEDER

Programme, Galicia 2013 Call.

REFERENCES

Albinsson, P.-A. and Zhai, S. (2003). High precision touch

screen interaction. In Proceedings of the SIGCHI

Conference on Human Factors in Computing Systems,

CHI ’03, pages 105–112. ACM.

Bae, S.-H., Balakrishnan, R., and Singh, K. (2008). ILoveS-

ketch: as-natural-as-possible sketching system for

creating 3d curve models. In Proceedings of the 21st

annual ACM symposium on User interface software

and technology, UIST ’08, pages 151–160. ACM.

Bederson, B. B. (2000). Fisheye menus. In Proceedings

of the 13th annual ACM symposium on User interface

software and technology, UIST ’00, pages 217–225.

ACM.

Benko, H., Wilson, A. D., and Baudisch, P. (2006). Pre-

cise selection techniques for multi-touch screens. In

Proceedings of the SIGCHI Conference on Human

Factors in Computing Systems, CHI ’06, pages 1263–

1272. ACM.

Dandekar, K., Raju, B. I., and Srinivasan, M. A. (2003).

3-d ﬁnite-element models of human and monkey ﬁn-

gertips to investigate the mechanics of tactile sense.

Journal of biomechanical engineering, 125(5):682–

691. PMID: 14618927.

Dietz, P. and Leigh, D. (2001). DiamondTouch: a multi-

user touch technology. In Proceedings of the 14th an-

nual ACM symposium on User interface software and

technology, UIST ’01, pages 219–226. ACM.

Hall, A. D., Cunningham, J. B., Roache, R. P., and Cox,

J. W. (1988). Factors affecting performance using

touch-entry systems: Tactual recognition ﬁelds and

system accuracy. Journal of Applied Psychology,

73(4):711–720.

Han, J. Y. (2005). Low-cost multi-touch sensing through

frustrated total internal reﬂection. In Proceedings of

the 18th annual ACM symposium on User interface

software and technology, UIST ’05, pages 115–118.

ACM.

Hornecker, E. (2008). ”I don’t understand it either, but it

is cool”; - visitor interactions with a multi-touch table

in a museum. In 3rd IEEE International Workshop

on Horizontal Interactive Human Computer Systems,

2008. TABLETOP 2008, pages 113–120.

Kabbash, P. and Buxton, W. A. S. (1995). The ”prince”

technique: Fitts’ law and selection using area cursors.

In Proceedings of the SIGCHI Conference on Human

Factors in Computing Systems, CHI ’95, pages 273–

279. ACM Press/Addison-Wesley Publishing Co.

Lewis, J. R., Potosnak, K. M., and Magyar, R. L. (1997).

Chapter 54 - keys and keyboards. In Marting, G. H.,

Thomas, K. L., Prasad V. PrabhuA2 Marting G. He-

lander, T. K. L., and Prasad, V. P., editors, Hand-

book of Human-Computer Interaction (Second Edi-

tion), pages 1285–1315. North-Holland, Amsterdam.

Lundstrom, C., Rydell, T., Forsell, C., Persson, A., and Yn-

nerman, A. (2011). Multi-touch table system for med-

ical visualization: Application to orthopedic surgery

planning. IEEE Transactions on Visualization and

Computer Graphics, 17(12):1775–1784.

Pinhanez, C., Kjeldsen, R., Tang, L., Levas, A., Podlaseck,

M., Sukaviriya, N., and Pingali, G. (2003). Creat-

ing touch-screens anywhere with interactive projected

displays. In Proceedings of the eleventh ACM interna-

tional conference on Multimedia, MULTIMEDIA ’03,

pages 460–461. ACM.

Potter, R. L., Weldon, L. J., and Shneiderman, B. (1988).

Improving the accuracy of touch screens: an experi-

mental evaluation of three strategies. In Proceedings

of the SIGCHI Conference on Human Factors in Com-

puting Systems, CHI ’88, pages 27–32. ACM.

Quevedo-Fern

andez, J. and Martens, J. B. (2012). Demon-

strating idAnimate: a multi-touch system for sketch-

ing and rapidly manipulating animations. In Pro-

ceedings of the 7th Nordic Conference on Human-

Computer Interaction: Making Sense Through De-

sign, NordiCHI ’12, pages 767–768. ACM.

Rekimoto, J. (2002). SmartSkin: an infrastructure for free-

hand manipulation on interactive surfaces. In Pro-

ceedings of the SIGCHI Conference on Human Fac-

tors in Computing Systems, CHI ’02, pages 113–120.

ACM.

Rodr

ıguez, I., G

omez-Meire, S., Barreiro, E., Rodeiro, J.,

and Campos, C. (2013). Sketch arm, custom clos-

ets rapid prototyping system. Procedia Computer Sci-

ence, 18(0):986–995.

Sears, A. and Shneiderman, B. (1991). High precision

touchscreens: design strategies and comparisons with

a mouse. Int. J. Man-Mach. Stud., 34(4):593–613.

Vogel, D. and Baudisch, P. (2007). Shift: a technique for

operating pen-based interfaces using touch. In Pro-

ceedings of the SIGCHI Conference on Human Fac-

tors in Computing Systems, CHI ’07, pages 657–666.

ACM.

Wang, F. and Ren, X. (2009). Empirical evaluation for ﬁn-

ger input properties in multi-touch interaction. In Pro-

ceedings of the SIGCHI Conference on Human Fac-

tors in Computing Systems, CHI ’09, pages 1063–

1072. ACM.

Worden, A., Walker, N., Bharat, K., and Hudson, S. (1997).

Making computers easier for older adults to use: area

cursors and sticky icons. In Proceedings of the ACM

SIGCHI Conference on Human factors in computing

systems, CHI ’97, pages 266–271. ACM.

Zhai, S., Hunter, M., and Smith, B. A. (2000). The metropo-

lis keyboard - an exploration of quantitative tech-

niques for virtual keyboard design. In Proceedings

of the 13th annual ACM symposium on User interface

software and technology, UIST ’00, pages 119–128.

ACM.

GRAPP2014-InternationalConferenceonComputerGraphicsTheoryandApplications

416