Stereoscopic Interactive Objects: Acquisition, Generation and

Evaluation

avio de Almeida e Silva

, Diogo Roberto Olsen

, Lucas Murbach Pierin

, Fl

avio Bortolozzi

and Edson Jos

e Rodrigues Justino

Programa de P

os-graduac¸

ao em Inform

atica - PPGIa, Pontif

ıcia Universidade Cat

olica do Paran

a - PUCPR,

Curitiba, Paran

a, Brazil

Centro Universit

ario Cesumar - Unicesumar, Maring

a, Paran

a, Brazil

Keywords: Learning Objects, Stereoscopy, Interactive Environments, Multidisciplinarity, Visualization.

Abstract:

Three-dimensional objects suffer from inaccuracies in the format of computational representation and artistic

drawing, losing visual acuity. On the other hand, photographs offer visual acuity, but they are not interactive,

that is, they do not offer the possibility of changing the perspective of visualization as a three-dimensional

model does. Thus, we have created a device capable of scanning a real object, allowing for the production of

interactive contents and visual acuity. In addition, the model created allows stereoscopic visualization, that is,

with a notion of depth. We produced a learning object from the scans of a real object, assessed by teachers and

students who work at medium, technical, and higher school levels. The interactive stereoscopic learning object

with visual acuity seems promising for the responses to a questionnaire and interviews showed it motivated

teachers and students due to its more realistic visualization and rich details. By allowing the visualization of

the piece from several angles, it can foster and satisfy curiosities.

1 INTRODUCTION

The study of artworks, anatomical pieces or botani-

cal items is just an example of visual analyses of real

objects common in scientiﬁc research, education, or

in the production system, among others. These visual

inspections are a means for solving issues such as mu-

seum restoration or quality control in the production

chain, among other applications.

However, it is not always possible, or even inter-

esting, for the observer to be in direct contact with the

item he studies. For example, a middle-level student

does not need to enter into a human anatomy labora-

tory, where there are chemical agents that make this

environment unhealthy, to study anatomy.

On the other hand, imprecise, inaccurate or static

models may not faithfully represent the objects to

be studied and hamper the teaching-learning process.

Research has been done to obtain visual anatomical

models that offer high visual acuity, such as the Vis-

ible Human Project (Ackerman, 1998), in which two

human bodies were sectioned with intervals of 1mm

and 0.33mm and these sections were photographed to

generate image bases of the whole human body.

In his paper, Ackerman describes that there are

two forms of image representation: those based on

photos and three-dimensional models (3D). Three-

dimensional models suffer from imprecision of both

artistic design and graphical representation, which are

based on geomorphic forms and mathematical for-

mulas to represent volumes and structures. These,

however, do not faithfully represent biological forms,

which are irregular in nature.

As for the content in photographs, Ackerman ar-

gues that a set of two-dimensional static images are

limited because they do not offer the possibility of

changing the view perspective of the photograph as

does manipulation of the real object.

According to the author, the understanding of

tridimensional structures is essential in many areas,

but the learning of them is a challenge. Photographs

are two-dimensional in nature and force the viewer to

a mental exercise of constructing a 3D visualization,

which can lead to inaccuracies in the understanding

of structures.

Besides anatomy, the same difﬁculties in under-

standing three-dimensional structures occur in other

areas of education, such as the study of artworks,

botany, design, among others. In order to overcome

these problems, a device that can scan real objects

Silva, F., Olsen, D., Pierin, L., Bortolozzi, F. and Justino, E.

Stereoscopic Interactive Objects: Acquisition, Generation and Evaluation.

DOI: 10.5220/0007796001650176

In Proceedings of the 11th International Conference on Computer Supported Education (CSEDU 2019), pages 165-176

ISBN: 978-989-758-367-4

165

for the generation of Learning Objects (LO) was de-

veloped. The generated LOs can be visualized on

different devices, such as screens, projectors, and

augmented reality, and include the notion of three-

dimensionality, depth, and volume, preserving visual

acuity, displaying details that are not visible to the

naked eye.

The educational contents produced in this research

have interactivity and stereoscopy (depth perception).

To these contents will be added data such as informa-

tion, deﬁnitions and characteristics present in the real

objects, which were scanned. It is believed that the

minipulation of an interactive and stereoscopic con-

tent can attract the interest of the students, contribut-

ing to the learning.

This paper introduces a solution for the digitiza-

tion of real objects that allows an interactive, stereo-

scopic visualization and preserves visual acuity (Sec-

tion 2). Section 3 addresses LOs and how the device

can be used for this purpose. Section 4 presents the

assessment of an LO generated from such digitiza-

tion. Section 5 contains our ﬁnal remarks.

2 FULL FRAMES

SEMI-SPHERICAL SCANNER

Considering Ackerman’s notes on the need for inter-

active digital models with ﬁdelity to real objects, we

worked out a method of digitization based on a set

of photographs with visual acuity, which enables the

generation of interactive and stereoscopic models. We

deﬁne visual acuity here as the possibility to visualize

in the digital model the same details found in the real

object, that is, preserving its color, shape, and tex-

tures.

Visual acuity is achieved through the appropriate

use of photographic techniques, correction of intrinsic

problems in the digital photography process, enhance-

ment of the acquired images (Krasula et al., 2017),

and the use of quality metrics for these images (Cheng

et al., 2017). Fidelity is also guaranteed for the acqui-

sition of the colors of the real object (Yu et al., 2016).

Moreover, according to Ackerman, for the under-

standing of three-dimensional structures, it is desir-

able that one can interact with the model. In this pa-

per, interaction is deﬁned as the possibility of observ-

ing the digital model from different angles, manipu-

lating and rotating it as if the object were in hands.

However, we used two-dimensional photographs;

in order to allow for a spatial and volume-based un-

derstanding of the scanned object, we used stereo-

scopic techniques to create depth perception in the vi-

sualization. On the other hand, the work of (Weber

et al., 2016) presents equipment for the acquisition of

photos with a camera, an arm and a turntable, where

the camera can be attached at different points in the

arm, which have vertical movement.

This process is completely manual and generates

problems in the acquisition for the generation of stere-

oscopy. In the works of (Solav, 2018) (Zhang et al.,

2017) (Xu et al., 2017) multiple ﬁxed cameras are

used for the acquisition, which may entail the need

for adjustments in the images, with the application of

geometric transformations due to distortions.

In order to enable the scanning of real objects in

accordance with the mentioned characteristics, we de-

veloped a Full Frame Semi-spherical Scanner - F2S2.

The term “semi-spherical scanner” refers to the fact

that the F2S2 takes pictures while moving the cam-

era such that it forms a semi-sphere around the ob-

ject, covering all the angles according to a predeﬁned

range. The term “Full Frames” refers to the photo-

graphic acquisition of all angles of the object. Figures

1(a) and 1(b) show F2S2.

(a) (b)

Figure 1: Full Frame Semi-spherical Scanner: (a) internal

view; (b) side view.

Until the visualization process is reached, three

steps are taken: initially, the photographs of all the

angles of the object being scanned are acquired; these

photos are then corrected, processed and organized

into a ﬁle called Stream2D; ﬁnally, Stream2D is used

to generate the visualization. Figure 2 shows the com-

plete process.

Figure 2: F2S2 framework.

CSEDU 2019 - 11th International Conference on Computer Supported Education

166

Acquisition

The acquisition process occurs by photographing

all the angles of the object in ultra-high deﬁnition

(UHD), 4k (3840 × 2160 pixels) or 8k (7680 × 4320

pixels) pictures. Scanning takes a predeﬁned angu-

lar range and a minimum scanning accuracy of 0.9

◦

into account. In order for the camera to acquire the

photographs, it is moved by a mechanical bracket that

has two axes (Axes: H – responsible for the horizon-

tal movement of the camera, and V - vertical move-

ment). The H and V axes move the camera along an

arc-shaped path with a predeﬁned radius and with the

object positioned in the center of the arc. Picture 3(a)

shows the motion and the axes.

The C axis tilts the camera so that it maintains the

focal point in the center of the sphere, that is, on the

object. In addition to the movement of the camera,

the object is positioned on a turntable that revolves

around its own axis (Axis B), so that it is possible to

obtain images from all sides of the object for each an-

gular position of the camera. The arc motion (angular

variation ϕ) and the object’s rotation (angular varia-

tion Θ) cause the camera to move in a semi-sphere

around the object, as observed in Figure 3(b). The

acquisition process can be seen in videos

(a) (b)

Figure 3: Acquisition process.

Processing

After the acquisition, the images are adjusted auto-

matically, starting with the correction - by software

- of imperfections intrinsic to the process of digital

photography, such as deformations caused to the im-

age by the optical set of the camera, removal of chro-

matic aberrations and the vignette. The colors of the

pictures are corrected to match the actual colors of

the object. These problems occur in all photographs,

digital or analogical, and are not usually perceived in

photographs of common use. For scientiﬁc and edu-

cational applications, however, aiming for greater ﬁ-

delity to the real object, these characteristics should

be corrected.

Acquisition: https://www.youtube.com/channel/

UCXEFzyZGrGlLANNrCawNgWQ anonymous link

Then, the context is reduced and the background

removed. These steps remove parts of the scanned

image, such as the inﬁnite background where the ob-

ject is positioned, so that the resulting image is the

scanned object only. Figure 4 shows the acquired im-

age (Figure 4(a)) and the ﬁnal image (Figure 4(b)).

(a) (b)

Figure 4: Image processing: (a) image as scanned by F2S2;

(b) ﬁnal image.

Finally, the Stream2D is created, in which the pro-

cessed images are organized so that for each image of

the set it is possible to know the position of the camera

during the scanning. The organization of Stream2D

is based on a matrix with two structures: Streams and

Frames. A Stream is a set of images for a given po-

sition of the camera, i.e., a set of photos with angular

variation Θ but without angular variation ϕ. A Frame

is a photograph from a certain point of view of the

object. Each Frame has a unique angular variation ϕ

and Θ.

Figure 5: Stream2D.

Considering a maximum scan accuracy of 0.9

◦

the number of Frames in a Stream is deﬁned by

0 < Frame ≤ 400, 400 being the multiplication of

the angular variation Θ by the scanning precision

(360

◦

÷ 0.9

◦

The number of Streams in a Stream2D is deﬁned

by 0 < Stream ≤ 100, 100 being the multiplication

of the angular variation ϕ, which is 90

◦

, by the scan

precision (90

◦

÷ 0.9

◦

). Hence, the number of Frames

in a Stream2D is 0 < Frame ≤ 40, 000(400Frames ×

100Streams).

Stereoscopic Interactive Objects: Acquisition, Generation and Evaluation

167

Visualization

Stream2D visualization takes place in graphics com-

puting environments using, among others, three-

dimensional Cartesian coordinate systems (X, Y, and

Z axes). In Cartesian coordinate systems, a point P is

deﬁned by a set of three values: P = (X,Y, Z), where

the values of each axis indicate the displacement on

the axis. Figure 6 shows this coordinate system.

Figure 6: Cartesian coordinate system.

Since F2S2 performs the acquisition process in a

semi-spherical format, the position of a given photo

can be deﬁned using a spherical coordinate system.

This coordinate system is also based on three dimen-

sions to deﬁne a position in space: P = (r, Θ, ϕ) (Fig-

ure 7), where:

r: distance from point P to the origin of the sys-

tem;

Θ: angular variation between the projection of

point P in the plane formed by the axes XY;

ϕ: angle formed by the line connecting P to the

origin and the Z axis.

F2S2, in turn, has a four-axis system (H, V , C and

B) for scanning objects (see Figure 8), where:

H: displacement of the camera on the horizontal

axis;

V : displacement of the camera on the vertical axis;

C: inclination of the camera in relation to the hor-

izontal axis;

B: rotation of the object on its own axis.

In the F2S2 coordinate system, the position of a

point (Frame) is given by F = (H,V, B).

The value of the C-axis does not have to be in-

formed to deﬁne the position of a Frame F, since, for

a certain position of H and V , there is only one in-

clination C that links F to the origin of the system.

C = arctangent(H,V ) calculates the value of C.

Figure 7: Spherical coordinate system.

Figure 8: The F2S2 coordinate system.

Since there are three coordinate systems involved

in the process of digitizing objects until their exhibi-

tion, one should establish a conversion system from

one to the other.

Considering the spherical coordinate system,

which represents the position of a Frame in relation to

the Object, this system can be converted to the Carte-

sian and F2S2 systems by Equations 1 and 2, respec-

tively.











X = r × sine(Θ) × cosine(ϕ)

Y = r × sine(Θ) × sine(ϕ)

Z = r × cosine(Θ)

(1)











H = r × cosine(Θ)

V = r × sine(Θ)

B = ϕ

C = Θ

(2)

To convert the Cartesian coordinate system used

in graphical computing environments to the spherical

CSEDU 2019 - 11th International Conference on Computer Supported Education

168

and F2S2 coordinate systems, Equations 3 and 4 are

used, respectively.











r =

+ Z

Θ = arccosine(Z ÷ r)

ϕ = arctangent(Y, X)

(3)











H = Z

V =

B = arctangent(Y, X)

C = arccosine(Z ÷ r)

(4)

The F2S2 coordinate system used in the acquisi-

tion of the images can be converted to the spherical

(Equation 5) and Cartesian coordinate systems (Equa-

tion 6).











r =

Θ = C

ϕ = B

(5)











X = r × sine(C) × cossine(B)

Y = r × sine(C) × sine(B)

Z = V

(6)

Since the coordinates in the F2S2 coordinate sys-

tem are known during the Frames acquisition process,

it is possible to calculate the spherical coordinates of

each Frame relative to the scanned object. Subse-

quently, to carry out the interactive visualization of

the generated model, the coordinates of the observer

are calculated in a graphical computer system (based

on the Cartesian model). The frame corresponding to

that display is found in the spherical model, and the

image associated with that position is displayed.

Interactivity

Interactivity during the visualization of the gener-

ated model may occur in two ways: navigation

and geometric transformations (rotation, translation,

and scale). Therefore, it is possible to manipulate

the images with the basic characteristics of a three-

dimensional model. Navigation allows you to change

the point of view to any angle of the object, that is,

changing the view from one position to another. Fig-

ure 9 shows a sequence with a frontal position, right

navigation, upward navigation, and, ﬁnally, naviga-

tion to a view from above.

Geometric transformations do not alter the point

of view of the object but lie over a photograph. Trans-

lation is the capability to move the image on the

screen, for instance, from the center to one of the cor-

ners. Rotating is changing the orientation of the im-

age, for instance, allowing a photograph with the ac-

quired object to be displayed in whatever inclination.

Scaling is the possibility of zooming the image in

or out, making it larger or smaller. As the acquisition

occurs in 4k or 8k, a non-pixelated view is obtained.

Normally, when you enlarge an image, say, by 800%,

it becomes blocky. But in an image with the pixel

density achieved by the F2S2, this does not happen.

Figure 10 shows a digitized model and a sequence

of changes delivered by these geometric transforma-

tions: 10(a) original image, 10(b) major scale, 10(c)

translation, and 10(d) rotation.

The interactivity of the generated model also al-

lows for the addition of hypermedia contents, buttons,

and regions of interest in the images, allowing, for ex-

ample, clicking on parts of the image to display infor-

mation about it, as shown in Section 4 - Figure 14(c).

Stereoscopy

The information on depth and proportion can be gen-

erated using some artiﬁcial resource. An example is

a single image captured by a common photo camera.

Depending on the type of image to be captured, the

notion of depth and proportionality is taken into ac-

count. (Murray, 1994) presents the deﬁnitions and

discusses the issue of perception of depth. One can

apply so-called passive effects, such as rotating by

◦

, to a static image. These effects are inherent to

the world as we see it and do not depend on the way

in which they are captured. The characteristics of

this type of effect are perspective, lighting, occlusion,

shadow, and texture gradient.

Perspective gives us the sensation of objects be-

ing larger or smaller than others depending on their

position. Illumination is an effect that contributes to

giving shapes to objects; it also produces the effect

of shadow, which is easy to visualize and understand,

as for shadow to exist there must be light. Occlusion

occurs when one object overlaps another, that is, it

stands in front of another hiding parts of the one be-

hind. In order to create a notion of depth, one can

also use texture gradient, since its effect is a repeti-

tion of the pattern present in the virtual environment

or object.

In short, all of these characteristics are related to

the 3D effect on virtually created objects, since they

are originally ﬂat, and that does not change with the

application of these effects. However, through their

application, one has the feeling that the object has

depth and proportion. You can see these effects in

photographs, as shown in Figure 11, without the need

for a software application.

Stereoscopic Interactive Objects: Acquisition, Generation and Evaluation

169

(a) (b) (c) (d)

Figure 9: Interactivity: navigation.

(a) (b)

Figure 10: Interactivity: geometric transformation.

Figure 11: Passive effects.

Unlike the types of passive effect cited, stere-

oscopy in humans is of the active type. This means

that two images are needed for the brain to form the

notion of depth and proportion. The brain treats an

image captured by a single eye as ﬂat, mainly lacking

the notion of depth. Stereoscopy comes from the fact

that the human vision presents a difference between

the images captured by the eyes due to their position

since, in humans, they are facing forward and with an

average separation of 6.5 cm. This difference, called

binocular disparity, allows for two slightly different

images, one for each eye, enabling the mentioned no-

tions, and known as the stereoscopic view.

As for the stereoscopic view on devices such as

TVs, monitors, cell phones or screen projectors, there

is also a concept of distance. However, it is related to

parallax. Parallax is understood to be the positioning

changes of an object in relation to the different points

of observation, or the apparent displacement of a ref-

erence from the movement of an observer. It allows

you to have a feeling of depth. In other words, the

stereoscopic view is obtained by means of the differ-

ence in the positioning of the eyes and the crossing

point of the images of each eye.

Parallax and Disparity for the formation of stereo-

scopic images are similar. As seen, parallax is related

to the projection plane of the images, so its measure-

ment is carried out in this plane. The disparity is a

retina-related measure, the distance between the eyes.

The use of some apparatus to view stereoscopic im-

ages, such as glasses, causes parallax to become the

so-called retinal disparity. Thus, parallax will pro-

duce retinal disparity and will be responsible for the

production of the stereo vision.

Parallax can be given in terms of angular measure,

thus relating it to disparity by taking the distance be-

tween an observer and the plane of projection into

account. As mentioned, there is a mean difference

(DM) between the eyes, given in centimeters; if you

also have the distance in centimeters between the ob-

server and the plane (d), you can calculate the angular

measure (α), as seen in Equation 7.

α = 2 arctan

(7)

Using the F2S2, you can create stereoscopy with-

out the need for transformation, which is very com-

mon when you make a picture with two parallel or

convergent cameras. In the latter case, the image must

be adjusted so that the common points of both the

right and left images are present; otherwise, you will

have problems with stereoscopy, which may cause

discomfort during visualization.

The F2S2 scanning process eliminates the need to

apply mathematical transformations, since it acquires

CSEDU 2019 - 11th International Conference on Computer Supported Education

170

the images in different angles, as happens in the hu-

man view, where each eye perceives the object from a

different position. Thus, the image scan enables stere-

oscopy by simply overlapping two scans.

The name of these two images is stereo pair.

Based on the parameters of parallax and disparity, the

selected images are 6

◦

apart (angular variation Θ).

Using a pair of images 2

◦

or 4

◦

apart generates less

intense stereoscopy. Pairs of images in which the dis-

tance is more than 6

◦

will cause discomfort.

The anaglyph technique was used for the stereo-

scopic visualization as seen in Figure 12. In this

technique, the color channels in the two images are

changed, placing the image seen by the left eye in one

color and the right one in another. Normally, the red

channels are used for the left images and cyan (blue

and green) images for the right ones, as was the case

in the ﬁrst few years of 3D cinema, although this de-

pends more on the colors present in the lenses of the

glasses. Despite the processing of the images, the

stereoscopic visualization maintains the original color

of the object.

Figure 12(b) was acquired with a 6

◦

difference in

relation to Figure 12(a) of which the green and blue

channels were removed, leaving the red channel only.

In Figure 12(b) the red channel was removed, leaving

the cyan. This process is accomplished at runtime by

means of Equation 8.

EA = 255 −

(255 − M) ∗ (255 − Ii)

255

(8)

AE (Anaglyph Stereoscopy) is the ﬁnal image, M

or Mask is the image of the upper level (Figure 12(a))

and Ii the image of the lower level (Figure 12(b)).

This is done with all the images acquired by F2S2,

obeying the criterion of angular variation Θ. For ex-

ample, the acquisition of a stream of 2 in 2 degrees,

obtaining 180 images, allows stereoscopy between

the 1

and 4

frames (0

◦

and 6

◦

, respectively), 2

and 5

◦

and 8

◦

), 3

and 6

◦

and 10

◦

), and

so on, repeating the angular variation Θ = 6

◦

. Since

the last are formed between the 177

and 180

(352

◦

and 358

◦

), 178

and 1

(354

◦

and 0

◦

), 179

and 2

(356

◦

and 2

◦

), 180

and 3

(358

◦

and 4

◦

), all possi-

ble visualizations of the object are covered.

3 LEARNING OBJECTS

Diversity within the educational environment fos-

ters discussion among all learners, both students and

teachers, about how to present content as they all try

to ﬁnd new ways of supporting the teaching-learning

process.

(a) (b)

(e)

Figure 12: Images: (a) Left, (b) Right, (c) Red, (d) Cyan

and (e) Anagliph.

We understand that it is in this context that Learn-

ing Objects (LO) come in (Redmond et al., 2018) (Sri-

vastava and Haider, 2017). A deﬁnition of LO, ac-

cording to the Learning Technology Standards Com-

mittee (LTSC-IEEE), is “any digital or non-digital en-

tities that can be used, reused or referenced during

learning with technological support”

Thus, we address here how to prepare the contents

generated by F2S2 for use in several domains, espe-

cially in education. That is, all the objects scanned

by F2S2 are considered “raw resources” ((Tarouco

et al., 2014) pg.17) for the creation of LOs, and should

be prepared for the devices of the visualization men-

tioned in section 2. In (Guterres and Silveira, 2017b)

LTSC - http://sites.ieee.org/sagroups-ltsc/home/

Stereoscopic Interactive Objects: Acquisition, Generation and Evaluation

171

is presented a framework that may help you develop

LOs.

From the point of view of education as dis-

cussed by (Wiley, 2002), the application of LOs it-

self does not guarantee that pedagogical objectives

are achieved. He questions the use of the LEGO

T M

blocks introduced by Hodigns (Hodgins, 2002) as an

LO, claiming it to be a simplifying metaphor, which

may neglect the challenges of creating educational ex-

periences. Despite this, due to the diversity within the

educational environment, every student’s need must

be taken into account. (Mend

ez et al., 2016) address

the concern of adequacy of each student to LOs.

A consensus is that LOs should contain reusable

contents, that is, they should meet this criterion for

their elaboration. Despite this deﬁnition, there are

barriers to the adoption of this type of initiative, since

institutions differ on what constitutes an LO. But

there are also various development problems, as seen

in (Guterres and Silveira, 2017a).

In an attempt to overcome these barriers, we seek

to develop reusable LOs in different areas of knowl-

edge that offer dynamic interaction, allowing partici-

pants to intervene freely to move an object to the de-

sired position, for use in educational environments,

for example, classrooms, or on their own devices such

as notebooks or Virtual Reality.

To be accessible and reusable, LOs are stored

in so-called repositories. An example is found in

(d. Silva et al., 2017). (Rodes-Paragarino et al., 2016)

address repositories adopted by teachers. Reposito-

ries make metadata available (LTSC, 2002) besides

ontologies, which also facilitate searches (Araujo,

2017) (Carvalho et al., 2017) (Lima et al., 2017)

(Sanches et al., 2017).

Repositories should provide LO metadata (MOA)

with a high degree of semantic interoperability. The

Open Archives Initiative Protocol for Metadata Har-

vesting (OAI-MPH)

, which is in version 2.0, aims to

collect metadata through an application-independent

interoperability framework.

However, this paper does not aim to produce

repositories nor metadata, but an LO with the objects

scanned by F2S2. This LO may be made available in

some repositories with metadata for access in the fu-

ture. The idea of this paper was to create an LO and

evaluate its feasibility through the participation of a

group of teachers from different areas and levels. We

describe the outcomes in section 4.

http://www.openarchives.org/OAI/2.0/

openarchivesprotocol.htm

4 LO ASSESSMENT

The objects scanned by F2S2 allow for the construc-

tion of an LO that offers interactivity and visual acu-

ity. However, it needs to be scrutinized by those who

might use it: students, teachers, and pedagogy pro-

fessionals. Therefore, we held two meetings to test an

LO with respect to acceptability.

In one meeting, we showed teachers and a profes-

sional in pedagogy an LO. In the other, a group of

students attended a lesson. After these meetings, we

handed out questionnaires with the objective of eval-

uating the LO and its acceptance by those involved.

We produced the LO from the digitization of a

human skull and added interactivity and stereoscopy

properties, as well as buttons identifying the bones.

When you click on a button, the skull rotates to the se-

lected bone, which appears on the screen and is high-

lighted. Figure 14 shows the LO screens used in the

search. Figure 14(c) shows the maxillary bone was

selected and evidenced.

Evaluation by the Teachers

The invited teachers work at different levels of educa-

tion, ranging from medium to higher levels. They re-

ceived manipulation instructions and used the tool as

if they were teaching a lesson. They were then asked

to comment on positive points or those that are ﬂawed

and should be improved.

We prepared the environment with a multimedia

projector and two notebooks with the tool running on

both. Thus the teachers could manipulate the tool on

the notebook connected to the projector to become

comfortable with the manipulation commands as well

as to view the object also on a monitor/TV and not

only projected on the wall/screen, for the projector

was not of high resolution.

This procedure was adopted for them to have a vi-

sual experience of quality that a TV/monitor, even if

only FullHD and not in 4k, could provide. Partici-

pants were able to discuss the use of the LO at the

levels in which they operate. They started to suggest

what would be interesting for a given level and what

could contribute to the learning process.

A cross-sectional study was carried out on May

28 and 29, 2018. Aiming for greater variability in the

sample, we invited 15 teachers working at the medium

technical level (computer science and biotechnology),

post-technical (oral and dental health), and higher

level (biological sciences and systems analysis and

development). The teachers were asked to analyze

if the same LO could be interesting for the differ-

ent courses. As the selection considered the courses

CSEDU 2019 - 11th International Conference on Computer Supported Education

172

Figure 13: Teachers’ answers.

(a) (b)

(c)

Figure 14: The Learning Object.

in which the participants work, the sample is non-

probabilistic.

All teachers signed a free and informed consent

form. After the presentation and effective use of the

tool, they answered a questionnaire with ten closed

questions, and a scale from Strongly Disagree to

Strongly Agree. The questions were:

1. Were you comfortable using the environment to

display the content to the students?

2. Do you think software like this can help you in the

teaching-learning process?

3. Would you use this environment in your classes?

4. Do you believe that you can use the environment

in a multidisciplinary and interdisciplinary way?

5. Do you believe that you can use the environment

for problematization?

6. Do you believe that you can use the environment

in active methodologies?

7. Do you think that this environment can be a way

of attracting the attention of the students?

8. Do you think that the present object, giving the

sensation of being an object similar to the real one

by its color, texture, and shape, can motivate the

students?

9. Can the manipulation of the object and visualiza-

tion of all angles encourage interaction/discussion

between the teacher and the students?

10. Can the manipulation of the object and the

visualization of all angles encourage interac-

tion/discussion among students?

We structured the questions in order to answer

three hypotheses based on Ackerman’s study, which

argues that the lack of interactivity (three-dimensional

visualization) and the imprecision of the models (vi-

sual acuity) hinder the learning process of anatomical

structures. The hypotheses were:

1. The content, with its interactive visualization fea-

tures and visual acuity, motivates the teacher in

the teaching-learning process. Questions 1 to 6;

2. The content, with its interactive visualization fea-

tures and visual acuity, motivates the student in

the teaching-learning process. Questions 7 and 8;

3. The interactivity and visual acuity of the envi-

ronment promote discussions between teacher and

student, and among students. Questions 9 and 10.

Hypotheses 1 and 2 were based on the premise

that the interactivity of the model can help understand

three-dimensional biological structures, and this can

motivate the teacher and the student. Hypothesis 3

Stereoscopic Interactive Objects: Acquisition, Generation and Evaluation

173

is based on the possibility that, in contrast to materi-

als such as a photo in a book, that directly exhibit a

particular structure, the interactive model enables ex-

amining the structure and visualizing it by different

angles. Interactivity was assessed in relation to pro-

moting in-class discussions when compared to static

materials. The answers obtained by the questionnaire

can be seen in Figure 13.

Questions 1, 2, and 3 address comfort and relia-

bility of the teachers’ use of the tool; comfort can be

determinant for its adoption in the classroom and can

help in the teaching-learning process. In response to

these questions, 93% of teachers indicated that they

agree or fully agree that they felt comfortable, believe

that the system can help to teach and that they would

use LO in the classroom. Two teachers were indiffer-

ent, and one did not answer, because he did not use

the tool but just watched the others.

Questions 4, 5, and 6 map the teachers’ believes

that this technology can be used in a multidisciplinary

and interdisciplinary way, for problematization or in

active methodologies. In response to these questions,

95% of the teachers said yes, one teacher did not re-

spond, and another was indifferent. As to the totality

of responses obtained on hypothesis 1, 94% of teach-

ers agree that the tool is motivational in the teaching-

learning process.

With regard to questions 7 and 8, linked to hy-

pothesis 2, all teachers agreed or fully agreed that LO

could be a motivating factor for students. Questions 9

and 10, related to hypothesis 3, outline if the interac-

tivity of the tool can promote discussions in the class-

room, since it encourages the student to search for the

information and not wait, solely and exclusively, for

the teacher’s answer. To this hypothesis, 96% of the

respondents agreed or fully agreed that the tool could

encourage in-room discussions. One teacher was in-

different.

In addition to the questionnaires answered by the

teachers, we invited a pedagogue to participate in

the presentation and later interviewed her in an open

interview, with predetermined questions and ample

freedom of reply. The teacher questioned the com-

putational resources needed to apply the technology:

“High-resolution TVs and expensive computers are

not a reality in most classrooms in Brazil”. In this re-

gard, in the coming years, these technologies tend to

become cheaper and become standard, thus entering

the classroom.

As for the present system, according to the peda-

gogue, the fact that the model is interactive and en-

courages the student to explore information makes

understanding of the subject of the lesson “complete”

and makes the tool a good option for active method-

ologies. Interactivity and the possibility of seeing all

the angles of the object can encourage interaction in

the class since it can make the student analyze the

whole object and not only the displayed angle. “It

can foster and engage curiosity”.

Based on the analysis of the answers, the LO, ac-

cording to the teachers, through its interactivity and

visual acuity, allowed for the visualization of details

that synthetic objects do not offer; nor do photos,

compared to this LO, allow full visualization and ma-

nipulation of the object.

Evaluation by the Students

In a similar way to what we did with the teachers, on

February 8, 2019, we taught a lesson to a group of 23

students of the technological course in Radiology, in

which we used the same LO presented to the teachers.

After the lesson, we gave the students a questionnaire

with the following questions:

1. Were you interested in the content because it was

in three dimensions?

2. Do you think software like this can help you in

your learning process?

3. Do you feel that this interactive environ-

ment might draw more of your attention than

static environments such as photos, books, and

slideshows?

4. The fact that the skull can be rotated to view other

sides can favor discussions about the topic of the

lesson?

5. Zooming in on more detail can foster debates on

the subject of the lesson?

6. Compared to drawings, photos, books, or

slideshows, can three-dimensional content en-

courage class discussion?

7. Do you think that the object presented gives you

the sensation it of being an object similar to the

real one by its color, texture and form?

8. Describe the experience of using the environment

and think of the possibilities it may offer to your

learning process. (Open-ended question).

The purpose of these questions was to map two

hypotheses:

1. The environment, with its three-dimensional vi-

sualization characteristics, interactivity, UHD and

ﬁdelity of form, color, and texture, motivates the

student in his learning process. Questions 1, 2, 3,

and 7.

2. The interactivity and ﬁdelity of the environment

promote discussions between teacher and students

and among students. Questions 4, 5 and 6.

CSEDU 2019 - 11th International Conference on Computer Supported Education

174

Figure 15: Students’ answers.

Hypothesis 1 is based on the premise that the inter-

activity and visual acuity of the technology used can

help in the understanding of three-dimensional struc-

tures, motivating the student in his learning process.

Hypothesis 2 examines whether interactivity can pro-

mote classroom discussions as it offers the student the

opportunity to search for the information instead of

readily getting it, just as in a book that shows the ideal

perspective for analysis of a given structure. Figure

15 shows the answers obtained by the questionnaires.

As for the answers, the perception of the stu-

dents involved in the study is that the characteristics

of interactivity, three-dimensionality, high resolution,

and ﬁdelity motivate the student in his learning pro-

cess since to questions 1, 2, and 7, 100% of the stu-

dents answered afﬁrmatively. Regarding question 3,

two students classiﬁed it as indifferent, and 91% re-

sponded afﬁrmatively.

According to the students’ response, the presented

technologies are able to motivate discussions in the

classroom since questions 4 and 6 received a 100%

afﬁrmative answers, and only one answer to question

5 was indifferent.

With regards to question 8, discursive, 15 of the

23 students answered the question. One reported: “...

I have difﬁculty in concentrating on my learning, and

these images helped me focus on the whole explana-

tion ...”. Other 08 students stated that the possibility

of visualizing structures in high resolution, with vi-

sual acuity and three-dimensionality could help them

in their learning process.

One of the students still wrote: “The presence of

the object in 3D gives a real notion of the mass and

shape of the object, which is a better way of under-

standing the function of the object”. This answer con-

ﬁrms Ackerman’s claim that interactivity and visual

acuity are important for the understanding of three-

dimensional structures.

5 CONCLUSION

This paper presented an evaluation of an LO produced

from scans of the Full Frames Semi-spherical Scan-

ner. As a contribution, the scanning method offers a

feature for a new form of 3D visualization with visual

acuity.

So, we created an LO by digitizing a human skull,

adding interactivity to the 3D visualization. This ob-

ject was manipulated by students and teachers, who

answered a questionnaire about the use of the tool.

The results show that interactive material is an alter-

native that can motivate teachers and students, as well

as being able to promote debates between teacher and

students and among the students.

Although the number of teachers who answered

the questionnaire is not sufﬁcient to generalize con-

clusions about the technology developed and the ob-

ject generated, those who manipulated the tool con-

sidered that the visualization adequately represents

the real object, and it can help in the teaching-learning

process.

Our next steps will involve the preparation of

learning objects to be used in class, and experiments

to evaluate the effectiveness of the objects compared

to other technologies without interactivity, visual acu-

ity, and stereoscopy.

REFERENCES

Ackerman, M. J. (1998). The visible human project. Pro-

ceedings of the IEEE, 86(3):504–511.

Araujo, A. (2017). Experi

encias l

ogico-matem

atica: um

objeto de aprendizagem de uso integrado de recursos

manipulativos digitais. (Cbie):546.

Carvalho, V., Ara

ujo, R. D., Ferreira, H., Cattelan, R.,

and Dorc¸a, F. (2017). OntAES: Uma Ontologia para

Sistemas Adaptativos Educacionais Baseado em Ob-

Stereoscopic Interactive Objects: Acquisition, Generation and Evaluation

175

jetos de Aprendizagem e Estilos de Aprendizagem.

(Cbie):1307.

Cheng, Y., Pedersen, M., and Chen, G. (2017). Evaluation

of image quality metrics for sharpness enhancement.

In Proceedings of the 10th International Symposium

on Image and Signal Processing and Analysis, pages

115–120.

d. Silva, J. W. F., d. Souza, C. T., and d. F. C. d. Souza,

M. (2017). Clover: An optimized repository for cus-

tomizable learning objects. In 2017 IEEE 17th Inter-

national Conference on Advanced Learning Technolo-

gies (ICALT), pages 76–78.

Guterres, J. and Silveira, M. (2017a). As Principais Diﬁcul-

dades Encontradas Durante o Processo de Produc¸

de Objetos de Aprendizagem. (Cbie):294.

Guterres, J. and Silveira, M. (2017b). Um Framework para

Apoio

a Reﬂex

ao sobre o Processo de Produc¸

ao de

Objetos de Aprendizagem. (Cbie):284.

Hodgins, H. W. (2002). The future of learning ob-

jects. http://www.coe.gatech.edu/eTEE and http:

//dc.engconﬁntl.org/etechnologies/11/ acessado em

2018/06/08.

Krasula, L., Callet, P. L., Fliegel, K., and Kl

ıma, M. (2017).

Quality assessment of sharpened images: Challenges,

methodology, and objective metrics. IEEE Transac-

tions on Image Processing, 26(3):1496–1508.

Lima, R. W. D., Oliveira, A. M. D., and Silva, P. D. A.

(2017). Ontologia para o Gerenciamento de Objetos

de Aprendizagem. (Cbie):1796–1798.

LTSC, L. T. S. C. (2002). Ieee standard for learning object

metadata. IEEE Std 1484.12.1-2002, pages 1–40.

Mend

ez, N. D., Morales, V. T., and Vicari, R. M. (2016).

Learning Object Metadata Mapping with Learning

Styles as a Strategy for Improving Usability of Educa-

tional Resource Repositories. Revista Iberoamericana

de Tecnologias del Aprendizaje, 11(2):101–106.

Murray, J. (1994). Some perspectives on visual depth

perception. ACM SIGGRAPH Computer Graphics,

28(2):155–157.

Redmond, C., Davies, C., Cornally, D., Adam, E., Daly, O.,

Fegan, M., and O’Toole, M. (2018). Using reusable

learning objects (RLOs) in wound care education: Un-

dergraduate student nurse’s evaluation of their learn-

ing gain. Nurse Education Today, 60(November

2016):3–10.

Rodes-Paragarino, V., Gewerc-Barujel, A., and Llamas-

Nistal, M. (2016). Use of Repositories of Digital

Educational Resources: State-of-the-Art Review. Re-

vista Iberoamericana de Tecnologias del Aprendizaje,

11(2):73–78.

Sanches, L., Costa, L., Salvador, L. N., and Souza, M.

(2017). Anotac¸

ao sem

antica autom

atica de Objetos de

Aprendizagem Digitais: Um mapeamento sistem

atico

de literatura. (Cbie):31.

Solav, D. (2018). MultiDIC: an Open-Source Toolbox for

Multi-View 3D Digital Image Correlation. 6.

Srivastava, B. and Haider, M. T. U. (2017). Personalized

assessment model for alphabets learning with learning

objects in e-learning environment for dyslexia. Jour-

nal of King Saud University - Computer and Informa-

tion Sciences.

Tarouco, L. M. R., Costa, V. M. d.,

Avila, B. G., Bez, M. R.,

and Santos, E. F. d. (2014). Objetos de Aprendizagem:

Teoria e Pr

atica.

Weber, J., Faria, V. D., Teixeira, M. J., Otoch, J. P., and

Figueiredo, E. G. (2016). Virtual and stereoscopic

anatomy: when virtual reality meets medical educa-

tion. 125(November):1105–1111.

Wiley, D. A. (2002). Connecting learning objects to in-

structional design theory: A deﬁnition, a metaphor,

and a taxonomy. http://reusability.org/read/chapters/

wiley.doc acessado em 2018/06/08.

Xu, G., Chen, J., and Li, X. (2017). 3-D Reconstruction

of Binocular Vision Using Distance Objective Gener-

ated From Two Pairs of Skew Projection Lines. IEEE

Access, 5:27272–27280.

Yu, H., Cao, T., Li, B., Dong, R., and Zhou, H. (2016). A

method for color calibration based on simulated an-

nealing optimization. In 2016 3rd International Con-

ference on Information Science and Control Engineer-

ing (ICISCE), pages 54–58.

Zhang, T., Liu, J., Liu, S., Tang, C., and Jin, P. (2017). A 3D

reconstruction method for pipeline inspection based

on multi-vision. Measurement: Journal of the Inter-

national Measurement Confederation, 98:35–48.

CSEDU 2019 - 11th International Conference on Computer Supported Education

176