Principles of 3D Web-collections Visualization

I. N. Sobolevskaya

and A. N. Sotnikov

Joint SuperComputer Center of the Russian Academy of Sciences, Branch of Federal State Institution “Scientific Research

Institute for System Analysis of the Russian Academy of Sciences”, 119334 Moscow, Leninsky Prospect, 32a, Russian

Keywords: Photogrammetry, 3D-modeling, Interactive Animation, Web Design, Polygonal Modeling.

Abstract: The paper introduces to approaches to solving the problem of creating realistic interactive 3D-web collections

of museum exhibits. The article deals with the representation of 3D-models objects based on oriented

polygonal structures and methods for 3D-models development based on photogrammetry, interactive videos,

and editable surface method. The evaluation of the computational complexity of constructing realistic 3D-

models is analyzed too. The calculation results for real museum exhibits are given. The paper describes

approaches to the formation of digital collections in the integration environment of the digital library. One of

the ways to form a multifunctional information resource is proposed as an effective solution to the problem

of presenting digital collections of various memory institutions to general public users. One of such methods

of forming a multifunctional information resource is implemented in the form of a virtual exhibition. The

format of virtual exhibitions allows combining the resources of partners to provide general public users of

collections stored in a museum, archive and library collections.

1 INTRODUCTION

In connection with the qualitative performance of

modern data transmission networks growth, new

opportunities are opening up for the information

systems operating organization with a large amount

of online information. Such information systems

include digital libraries (DL), in the collections are

represented digital objects of various origin (for

example, print publications, multimedia objects). If

several years ago a digital library's collection

included just only text files were accepted by the

achievement, then in modern realities collections of

document scans and digital 3D-models of museum

collections are already qualified normally.

To get to a whole new level of presenting objects

from archives and museums in the digital library as

three-dimensional objects it is necessary to develop

existing approaches and methods for obtaining 3D-

models, as well as create new ones.

However, in capture museum objects and

integrating them into a united information resource

are arise some difficulties. Such as:

- the creation of high-quality (in the context of visual

sensation) digital 3D-models of museum objects;

https://orcid.org/0000-0002-9461-3750

https://orcid.org/0000-0002-0137-1255

- the development of methods for describing

information objects which provide easy access to

them;

- the integration of digital 3D-models of museum

objects into thematic collections;

- the drill-down of information objects into the digital

library environment to create interdisciplinary

multimedia digital collections and virtual exhibitions.

At current, there is a limited supply of methods of

digitizing museum objects. Particularly as regards the

building of digital 3D-models and integrating them

into a united information resource, and the providing

digital 3D-models general public users with modern

multimedia technologies. This is particularly true in

science museums activities. For example for

organizing subject exhibitions, research practice and

supporting the academic programs.

These days there are two technologies used when

3D scanning museum objects: active and passive

based on contact and contactless 3D scanning,

respectively.

The use of these 3D scanning technologies allows

creating a high-quality three-dimensional digital

model of a museum object.

Sobolevskaya, I. and Sotnikov, A.

Principles of 3D Web-collections Visualization.

DOI: 10.5220/0008184401450151

In Proceedings of the 3rd International Conference on Computer-Human Interaction Research and Applications (CHIRA 2019), pages 145-151

ISBN: 978-989-758-376-6

145

The main shortcomings of the active scanning

method are the processing complexity and the

necessity to use high-resolution scanners when

scanning small-sized objects. The main shortcomings

of the contactless passive scanning method are the

inability to scan the "deepenings" of an object or its

defects, etc.

There is a section called "Virtual Exhibitions" on

the many museums' websites. Each museum defines

and interprets the concept of "virtual exhibition" by

itself. Some museums place hundreds of printed

and/or archival digital copies of documents in this

section, other museums offer a full catalog created

according to all the rules of multimedia technologies

as a "virtual exhibition". Work on the creation and

presentation of digital 3D-models museum objects

has been conducted since the 1990s of the 20th

century. In (Higuera, 1999; Hazan, 1999; Rao, 1999;

Zheng, 1998; Johnson, 1997; Nicchiotti, 1997;

Terashima, 1994) are described in detail methods for

the formation and presentation of 3D-models

museum objects in the digital space. The most

difficulties in creating a digital 3D-model arise from

the need to use large computational powers in the

calculations. In turn, to represent the constructed

model, specialized visualization software is required.

The development of multimedia and interactive

technologies allows today:

1. To demonstrate the museum objects that are not

included in the main exhibition of the museum (most

of the museum items are in the museum storage).

2. To form a virtual museum content, significantly

expanding the natural-science and cultural-

educational knowledge space;

3. To provide the possibility of obtaining more

detailed and more vivid (in an emotional sense)

information about the object, phenomenon, historical

event;

4. Turn the modern museum space into an educational

and leisure center.

Thus, if multimedia technologies are a well-

thought-out methodology of informational and

educational nature the introduction of these

technologies is not only justified but also is necessary

in some cases. At the same time, not every science

museum can turn its space into an entertainment and

leisure center. In this case, it is possible to create a

virtual exhibition space. Such a virtual exhibition can

be created together with archives and libraries. Then

these exhibitions could be built up on a digital library

platform. A technological and fundamentally new

resource is the digital library "Scientific Heritage of

Russia" (DL SHR) [http://e-heritage.ru/index.html].

DL SHR allows solving the problems of creating and

support of interdisciplinary virtual collections. DL

SHR is implementing the open remote access to its

funds for a general public user which significantly

distinguishes this data resource from many other

multimedia projects (Sobolevskaya, 2017).

At present, there is no common methodology for

applying one or another computer-generated

simulation method for storing the obtained volume

models with the possibility of subsequent replication

and providing users with the opportunity to inspect

virtual copies in various projections. The

development of a unified scheme for the application

of computer-generated simulation methods for

building 3D-models to create thematic 3D-collections

of museum objects in the digital library environment

and create thematic interdisciplinary virtual

exhibitions is relevant.

The analysis and development of methods for

generating 3D-models in the direction for obtaining a

realistic representation of 3D-collections in the Web-

space are open up opportunities for the formation of

high-quality 3D-collections of museum objects.

These methods allow to keep safe the originals and to

increase access to the high-quality digital copies of

museum exhibits.

2 MANUSCRIPT PREPARATION

Attempts to provide more complete information

about the object than a simple graphic image began

with the creation of stereo image technology.

The stereo image technology is based on the

representation of the object images taken by cameras

from different positions. With the advent of computer

graphics technology, the development of this idea

made a quantum leap. It became possible to store and

present an object image not only from two fixed

points but also from any point defined by the user.

It is clear, complete high-quality and detailed

digital 3D-modeling of an object is a rather difficult

task. This task needs the use of powerful computing

resources and special equipment. Such modeling

opens up new opportunities for the realizing of 3D-

models, their presentation, and use. For example, if

you have a high-quality 3D model built by dint of a

3D printer, you can get a copy of a museum object

halfway around the world without risking damage or

loss of the main artifact. This technology was used by

the State Darwin Museum of Moscow to collection

development with unique objects of digitized

archetypes of the skeleton elements of a prehistoric

man from the collection of the museum located in

South Africa.

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

146

To map a three-dimensional object on a plane

screen, you must project into the plane the three-

dimensional object. This is possible using projection

rays connecting the center of the projection, each

point of the object. These projection rays are required

to be orthogonal to the projection plane. This plane is

called the pictorial surface. It is located between the

displayed object and the observer, perpendicular to

the visual direction. The points at which the

projecting rays intersect the pictorial surface are the

corresponding points of the projection.

The main method of representing 3D-models

objects is polygonal modeling at the moment

(Lyashkov, 2017). The principle of method consists

in representing the object surface using the

coordinates of points on the object surface and the

normal vectors in them. This representation makes it

possible to approximate the reconstructed surface

with polygons with minimum error. In terms of the

linear algebra, the 3D model is represented as related

sets of nonplanar oriented triangles or more polygons.

Affine space and a homogeneous coordinate system

are used to represent elements of 3D objects.

The homogeneous coordinate system has the

property that the object defined in these coordinates

does not change when the same non-zero number

multiplies all coordinates. The number of coordinates

required to represent points is always one more than

the space dimension in which these coordinates are

used. For example, to represent a point in two-

dimensional space you need 3 coordinates and you

need 4 coordinates to represent a point in three-

dimensional space (Weeger, 2018).

Thus, for example, for manipulating 3D models

built in homogeneous coordinates to perform atomic

operations such as rotation, elation, scaling is

required square 4-order matrices. These matrices are

applied to all elements of the model.

The polygons geometry and their quantity are the

main parameters that determine the quality of a 3D

object model. There are two approaches to

determining the values of these parameters. Choosing

an approach mainly depend on the computational

power of the processors involved in drawing the

models. The first approach is characterized by an

increase in the number of polygons of minimal

complexity (triangles), which is accompanied by an

increase in the amount of data for displaying a 3D

model.

The second approach is based on the construction

of complex polygons that best fit the surface of the

simulated object.

The visualization algorithm based on the

formation of the Spline surface to form surfaces of

non-analytical forms described by composed

functions is one of the much-used tools in CASE

systems and computer graphics programs. This is

since the Spline is a differentiable curve. It contains

vertices and segments, passes through two

watchpoints an at least, and having watchpoints that

are not on this curve. These points determine the

Spline shape. At the same time, each Spline vertex

has a surface tangent vector passing through the

watchpoints. Each surface tangent vector has a so-

called marker. This marker controls the curvature of

the Spline segments at the entrance to the vertex. This

surface tangent vector belongs to the same vertex at

the exit from it. The most common types of splines

used in 3D graphics are Bezier curves and B-splines.

The use of splines in the methods of constructing 3D-

models allows to increase the quality of the simulated

surface, however, increases the complexity of the

calculation, which requires large computational

powers.

The construction of 3D models from a set of

images by photogrammetry is a rather laborious task

and requires rather bid computation power. For

example, 124 photos of image processing at one of

the cluster node MVS-10P (JSCC RAS) (Sotnikov,

2018) took 41 hours of calculations. Figure 1 shows a

digital 3D-model of anthropological reconstruction

constructed by photogrammetry by M.M. Gerasimov

located in the State Biological Museum named K.A.

Timiryazev.

Figure 1: 3D model produced by photogrammetry.

The limitations of this method are:

- it is impossible to control the process of creating a

3D model;

- waiting for long-running operations for the

processing of source data;

- need for sufficiently large computational powers.

Principles of 3D Web-collections Visualization

147

Thus the problem of creating high-quality 3D models

from a set of “flat” images requires solving the

following tasks:

• maximizing performance of data processing. Both

by increasing the powerful computing tools and the

creation of effective algorithms and programs;

• reduction of model setup time.

These tasks remain relevant and their solutions

involve the further development of methods for

digital visualization of objects.

Modern video-processors can process a stream of

data to draw graphical primitives. Modern graphics

processors in addition to improved performance

allow real-time execution of smoothing algorithm,

approximation, and optimization of polygons, which

leads to a decrease in the data volume of the 3D model

and improved visual perception.

To create a virtual collection of 3D-models to

provide it to general public users via the Internet is

used the so-called interactive animation technology.

This technology does not imply the building a full-

fledged 3D model but is based on a programmatic

change (scrolling) of a fixed set of object types

(frames) using specialized interactive mapping

programs that simulate a change in the point of view

of the original object. To create such an interactive

animation you need a set of pre-shot scenes that will

be used as separate exposure frames (Figure. 2).

Figure 2: Images to create an interactive animation.

The technology of interactive animation was

used, for example, in such projects as “Interactive

walk around the Kremlin” or “Virtual walk through

the garden of life”, dedicated to the 160th anniversary

of I.V. Michurin (Kirillov, 2017).

Based on the same data set (individual exposure

frames), a full-fledged 3D model can be constructed

using photogrammetry methods, which were actively

developed in the 70s for building a topographic map

using space and aerial photographs. Photogrammetry

uses methods and techniques of various disciplines,

mainly borrowed from optics and projective

geometry (Jo, 2017).

In the simplest case, the volume coordinates of

object points are determined by measurements taken

from two or more photographs taken from different

positions. The main task, in this case, is to determine

common points on two adjacent images. After

creating an array of the seed point cloud is formed a

set of rays. These rays pass through each common

point and through the location point of shooting

camera. The cross-cup of these rays determines the

location of a point on the surface of the original object

in space. More complex algorithms can use other,

known in advance, information about the object: for

example, the symmetry of the object's elements,

which, in some cases, allows you to reconstruct the

volume coordinates of object points from only one

photographic image.

The algorithms used in photogrammetry are

intended to minimize errors come about from

obtaining a set of consecutive frames, as well as

measurement errors. The Levenberg – Marquardt

algorithm (Xu, 2018), usually solves the problem of

minimizing the set of errors. This algorithm is based

on nonlinear systems solving by the least-squares

solution.

Objects that we encounter in life can be divided

into simple and complex. A simple object is an object

that can be described by a given (described by a given

equation) surface of revolution not higher than cubic.

An example of a simple object is parallelepiped,

which is a brick or a cylinder, for example. A complex

object is an object consisting of a combination of

simple objects. An example of a complex object is

almost any three-dimensional museum piece, such as

a stuffed mammoth, or an archaeological finding in

the form of an amphora.

Except for the described above some methods of

modeling in three-dimensional graphics, working

with 3D modeling is relevant to working with editable

surfaces (Hernando, 2018).

There are several ways to create 3D models using

editable surfaces.

These methods are based on the formation of a

model consisting of:

1. Triangular faces. In this case, the variable

parameters are the vertices of the triangles, their sides

(edges), the entire surface of the triangle (used, for

example, to overlay/change the texture) (Bowick,

1995);

2. Polygons. In this case, the variable parameters are

the vertices of the polygons, their sides (edges), the

entire surface of the polygon (Meilapredoviciu, 1991;

Kulikajevas, 2019);

3. Combinations of triangular or quadrangular faces

that are created by Bezier Splines (Schulz, 2011;

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

148

Barsanti, 2017). The two-dimensional Bézier Spline

is defined by four points: two endpoints to which the

ends of the curve are attached, and two control points

that act as "magnets" for pulling the curve away from

the line connecting the two extreme points (Dimas,

1999).

When you create 3D models in practice, you must

have to have the ability to move the control points

("sewing" of elementarily symmetric functions) by

yourself on the computer screen. The peculiarity of

this type of editable surface is the flexibility of

controlling the shape of the object being created. This

explains the usability of Bezier splines (Hu, 2017;

Zimmermann, 2017);

4. Objects created using Boolean operations. Boolean

operation is designed to create a new object based on

two or more existing objects. As a result of applying

the Boolean operations, a new object is formed as a

combination of the original objects. The combining

of the original object can be realized by addition

operation, subtraction operation and intersection

operation. The increase in the number of Boolean

operators injections in one another leads to an

increase in computational costs. Therefore it is almost

impossible to create a three-dimensional model of a

human head (especially a face) using this method

(Wu, 2018; Sheng, 2018).

When forming a visual image of a 3D object it

should not be overlooked the geometry, texture, and

features of the source object. Each object can be

modeled in several ways. When choosing how to

create a digital model of a 3D object should not be

overlooked several basic parameters: the parameter of

the object creation time and resources consumption

(the memory required to create and store the model).

Therefore, one of the main tasks of the formation of

3D digital models is the task of optimizing the

processes of modeling and visualizing objects. The

method of photogrammetry makes it possible to build

a high-quality 3D model with lower hard costs for

added hardware in comparison with other

visualization methods.

3 CONCLUSIONS

Based on the analysis and experiments using various

scanning equipment, the advantages and

disadvantages of various methods for constructing 3D

models were identified. Namely:

- existing methods of creating a 3D-model of a real

object have limitations in the construction of a

full-fledged realistic three-dimensional model of

an object that has a transparency surface, specular

surface or light-absorbing surface;

- To build 3D-models with a light-reflecting or

light-absorbing surface it is necessary to use

special equipment. Namely, it is necessary to use

a laser projector with a backlight grid and

operating in such a wavelength range in which the

“problematic” surfaces of the scanned object are

opaque. Such a source of illumination of the

reference grid should have a tuned working

frequency for working with objects from various

materials;

- Modern 3D-scanners are not allowed to obtain a

polygonal mesh for building a 3D- model of a

"transparent" object without applying a special

"marking" means that absorbs light onto its

surface. Such a scanning method is highly

undesirable, for example, for museum objects;

- Photogrammetry method requires very significant

computation efforts. Most of these efforts are paid

and set on cloud storages are existed on a third-

party file server. In addition to the method of

photogrammetry, in particular, implies long-

running operations for processing the source data

when creating a 3D model, as well as the use of

sufficiently large computing power.

When creating a digital 3D-model of a museum

object the main task is its complete safety

Based on the research and findings, the

technology for generating digital 3D-models of

museum objects using photogrammetry was taken as

the basis since this technology allows building a full-

fledged 3D model of an object with a complex

structure without using any special tools applied to

the surface of the object being photographed. For

visualization and presentation of a virtual collection

of 3D-models of museum objects to general public

users, the use of interactive animation technology is

proposed.

Digital 3D-model is an object that can be linked

by mutual links with other types of digital objects

likes of text, archival, audio and video documents,

etc. These links are stored on remote servers that can

be accessed via the Internet. A multimedia object,

which includes a set of digital 3D-models additional

information associated with them can be integrated

into some generic knowledge space. One of the

effective means of integration and presentation of

such information objects is the digital library.

One of the possible ways to solve the problem of

presenting digital museum collections and, in

particular, digital 3D-models of museum objects to

general public users via the Internet offers a method

Principles of 3D Web-collections Visualization

149

of creating a virtual exhibition immersed in a digital

library environment.

The results formed the basis of the technology of

creating 3D-models for objects from the funds of the

State Biological Museum named K.A. Timiryazev

(GBMT) and the formation on their basis by means of

the digital library "Scientific Heritage of Russia" of a

virtual exhibition dedicated to the scientific activities

of M. M. Gerasimov and his anthropological

reconstructions [http://acadlib.ru/].

During the production process on the virtual

project the collection of 3D-models of M.M.

Gerasimov's anthropological reconstructions was

created. Then this digital collection was downloaded

into the environment of the DL SHR. The virtual

exhibition is an example of intended use the

integration of libraries, archives and museums

resources into a unified thematic project.

The logic of virtualization of this exhibition is

based on building links between the person (M.M.

Gerasimov) to all elements of DL SHR, including

with museum objects and collections.

This virtual exhibition combines the resources of

libraries, archives, and a museum. In particular, this

project offers the visitor to get acquainted with the

history of anthropology, digitally printed publications

on anthropology, photo and video documents related

to the development of anthropology and

anthropological reconstructions, as well as 3D models

of M.M. Gerasimov's anthropological

reconstructions, stored in the GBMT funds.

The research is carried out by Joint

SuperComputer Center of the Russian Academy of

Sciences – Branch of Federal State Institution

“Scientific Research Institute for System Analysis of

the Russian Academy of Sciences” within the

framework of the State task 0065-2019-0014.

REFERENCES

Barsanti, SG., Guidi, G. (2017). A geometric processing

workflow for transforming reality-based 3d models in

volumetric meshes suitable for fea. In 3d virtual

reconstruction and visualization of complex

architectures, vol. 42-2, pages 331-338.

Bowick, M., Difrancesco, P., Golinelli, O., Guitter, E.

(1995). In Nuclear physics B, vol. 450(3), pages 463-

494.

Dimas, E., Briassoulis, D. (1999). 3D geometric modelling

based on NURBS: a review. In Advances in

engineering software, vol. 30(9-11), pages 741-751.

Hazan, S. (1999). Linking and thinking:

the Museum@School. In Proceedings of International

Conference on Museums and the Web, pages 59-65.

New Orleans, LA.

Hernando, A., Bobadilla, J., Ortega, F., Gutiérrez, A.

(2018). Method to interactively visualize and navigate

related information. In Expert Systems with

Applications, vol. 111, pages 61-75.

Higuera, JFD., Pernas, FJD. (1999). MINERVA:

Multimedia on the internet for virtual arts. In

Proceedings of 6th International Conference on

Multimedia Computing and Systems, vol. 2, pages 825-

829. Florence, Italy.

Hu, G., Cao, H., Qin, X., Wang, X. (2017). Geometric

design and continuity conditions of developable

lambda-Bezier surfaces. In Advances in engineering

software, vol. 114, pages 235-245.

Jo, Y.H., Kim, J. Y. (2017). Three-dimensional digital

documentation of heritage sites using terrestrial laser

scanning and unmanned aerial vehicle

photogrammetry. In International Archives of the

Photogrammetry, Remote Sensing and Spatial

Information Sciences, vol. 26, pages 395-398.

Johnson, J. (1997). The virtual endeavor experiment: A

networked VR application. In Proceedings of 4th

International Conference on Hypermedia and

Interactivity in Museums, pages 174-180. Paris, France.

Kirillov, S.A., Sobolevskaya, I.N., Sotnikov, A.N. (2017).

Ispol'zovanie mul'timedijnyh tekhnologij pri

formirovanii virtual'nogo estestvennonauchnogo

muzejnogo prostranstva. In Informacionnoe

obespechenie nauki: novye tekhnologii, pages 201–207.

Kulikajevas, A., Maskeliunas, R., Damasevicius, R. (2019).

Reconstruction of 3D Object Shape Using Hybrid

Modular Neural Network Architecture Trained

on 3D Models from ShapeNetCore Dataset. In

SENSORS, vol. 19(7), № 1553

Lyashkov, A.A., Panchuk, K.L, Varepo, L.G. (2017).

Osobennost' otobrazheniya giperpoverhnosti

chetyrekhmernogo prostranstva. In Geometriya i

grafika, pages 3-10. Moscow.

Meilapredoviciu, M. (1991). Cross sectioning algorithm for

solid modeling systems. In Computers and artificial

intelligence, vol. 10(5), pages 477-485

Nicchiotti, G. (1997). A system for 3D reconstruction and

VRML conversion for virtual museums applications. In

Proceedings of 13th International Conference on

Digital Signal Processing, pages 421-424. Santorini,

Greece.

Rao, AR.,Jaimes, A. (1999). Digital stereoscopic imaging.

In Proceedings of Conference on Stereoscopic Displays

and Virtual Reality Systems VI, pages 59-65. San Jose,

CA.

Schulz, T., Juttler, B. (2011). Envelope computation in the

plane by approximate implicitization. In Аpplicable

algebra in engineering communication and computing,

vol. 22(4), pages 265-288.

Sheng, B., Liu, B., Li, P., Fu, H., Ma, L., Wu, E. (2018).

Accelerated robust Boolean operations based on hybrid

representations. In Proceedings of 12th International

Conference on Geometric Modeling and Processing,

vol. 62, pages 133-153. Aachen, Germany.

CHIRA 2019 - 3rd International Conference on Computer-Human Interaction Research and Applications

150

Sobolevskaya, I.N., Kirillov, S.A. (2017). About the task of

creating a virtual museum space. In Software journal:

theory and applications, vol. 4, pages 1-3.

Sotnikov, AN., Sobolevskaya, I.N., Kirillov, S.A.,

Cherednichenko, I.N. (2018). Subject-oriented and

interdisciplinary digital collections in the electronic

environment knowledge. In Proceedings of XXth

Russian scientific conference “Scientific service on the

Internet”, pages 448-453. Novorossiysk, Russia.

Terashima, N. (1994). Telesensation - distributed

interactive virtual-reality - overview and prospects. In

Proceedings of IFIP 13th World Computer Congress,

vol. 51, pages 49-59. Hamburg, Germany.

Weeger, O., Dunn, M.S., Yeung, S.K. (2018). Fully

isogeometric modeling and analysis of nonlinear 3d

beams with spatially varying geometric and material

parameters. In Сomputer methods in applied mechanics

and engineering, vol. 32, pages 95-115.

Wu, Q., Xu, K., Wang, J. (2018). Constructing 3D CSG

Models from 3D Raw Point Clouds. In Computer

graphics forum, vol. 37(5), pages 221-232.

Xu, B., Cai, Y. (2018). A multiple-data-based efficient

global optimization algorithm and its parallel

implementation for automotive body design. In

Advances in Mechanical Engineering, vol. 10(8), №

1687814018794341.

Zheng, JY., Abe, N. (1998). Virtual recovery of excavated

archaeological finds. In Proceedings of International

Conference on Multimedia Computing and Systems,

pages 348-357. Austin, TX.

Zimmermann, Ch., Sauer, R.A. (2017). Adaptive local

surface refinement based on LR NURBS and its

application to contact. In Computational mechanics,

vol. 60(6), pages 1011-1031.

Principles of 3D Web-collections Visualization

151