06 April 2021
AperTO - Archivio Istituzionale Open Access dell'Università di Torino
Original Citation:
Exploring cultural heritage repositories with creative intelligence. The Labyrinth 3D system
Published version:
DOI:10.1016/j.entcom.2016.05.002
Terms of use:
Open Access
(Article begins on next page)
Anyone can freely access the full text of works made available as "Open Access". Works made available under a
Creative Commons license can be used according to the terms and conditions of said license. Use of all other works
requires consent of the right holder (author or publisher) if not exempted from copyright protection by the applicable law.
Availability:
This is a pre print version of the following article:
This version is available http://hdl.handle.net/2318/1578514 since 2017-12-04T09:32:41Z
Exploring cultural heritage repositories with creative
intelligence. The Labyrinth 3D system.
Rossana Damiano1,1, Vincenzo Lombardo1,1, Antonio Lieto1
a
Dipartimento di Informatica, Università di Torino
b
CIRMA, Università di Torino
Abstract
In cultural heritage, the use of ontologies makes the description of artworks
clearer and self-explanatory, with advantages in terms of interoperability. The
current shift towards semantic encoding opens the way to the creation of in-
terfaces that allow the users to build personal paths in heritage collections by
exploiting the relations over the artworks.
In the attempt to leverage this multiplicity of paths, we designed and imple-
mented a system, called Labyrinth 3D, which integrates the semantic annotation
of cultural objects with the interaction style of 3D games. The system immerses
the user into a virtual 3D labyrinth, where turning points and paths represent
the semantic relations over cultural objects, with the goal of engaging the user
in the exploration of the collection.
Keywords: 3D visualization, cultural heritage, computational ontologies
1. Introduction
In the last decade, the advent of connected, portable devices, and the evo-
lution of the Web towards a participatory model have prompted cultural insti-
tutions to pursue new communication strategies that leverage the Web [1, 2].
Cultural institutions have rushed to publish their collections online, with the
goal of innovating their interaction with the audience through the help of per-
sonalization and social media [3, 4].
In parallel with this trend, digital archives have moved towards semantic
annotation, a paradigm where the items in the archive are described with refer-
ence to a computational ontology. The use of ontologies, implemented through
logic-based languages [5], makes the description of artworks clearer and un-
ambiguous, with advantages in terms of interoperability among systems [6, 7].
Semantically annotated collections, then, lend themselves to personalization [4]
and cross media integration of data sources, following the paradigm of Linked
Open Data [8, 9].
Despite the potential of the semantic representation, however, the search in
heritage archives is still largely based on keywords and/or tags, through which
users can filter the archive contents to find what they need. As exemplified
Preprint submitted to Entertainment Computing June 30, 2016
by the well known Europeana initiative, which provides a unified interface to a
set of national digital collections [10], the search typically returns a list of items
(books, pictures, videos, etc.) accompanied with personalized recommendations,
but it does not contain an explicit representation of meaning relations over them.
In contrast with this approach, [11] argues that, in order to meet the needs of
the general audience, tools for supporting exploratory search are needed besides
the traditional keyword–based interfaces. In cultural heritage, search interfaces
are typically based on the metaphor of the “archive”, which mirrors the actual
fruition of the physical cultural objects (see, for example, the web interface of
the above mentioned Europeana system), although the trend of the 3D “visit”
has emerged in online museum collections, as demonstrated by the well known
Google Art Project.1
In this paper, we address the access to digital collections by proposing an
approach that leverages semantic annotation to create a 3D environment where
the user can explore the semantic relations over the items in a visual environ-
ment. Our approach combines the use of the 3D language, typical of new media
– and video games in particular –, with the capability of semantic annotation
to connect entities that are distant in space and time but share some common
features at the cultural level. The use of 3D for the interface is motivated by the
goal of attaining a user experience characterized by high level of engagement
and a sense of immersion [12]. As shown by an established line of research in
information visualization [13], in fact, visual metaphors can convey a conceptual
model in an immediate and engaging way.
The system we describe in this paper is part of a larger project, called
Labyrinth2, aimed at the dissemination of cultural heritage archives to the gen-
eral audience. In order to mediate between the point of view of the user and
the heterogeneity of the items in heritage repositories, which usually di↵er by
features such as media type, age and purpose but share some narrative features
like stories and characters, the project relies on the notion of “archetypes” of
narrative nature. Mainly inspired by the research in iconology and narratology
[14, 15], the term “archetype” is employed in Labyrinth to refer to a conceptual
core set at the intersection of narrative motifs, iconological themes and classical
mythology (the system itself is named after a well–known archetype).
The plan of the paper is the following: after describing the background of
the project and discussing its motivations (Section 2), in Section 3 we provide
a brief overview of the live system with a navigation example. Section 4 de-
scribes the components of the systems, namely the ontology (Section 4.1), the
3D environment (Section 4.2) and the core component of the system, i.e., the
mapping of the ontology onto the 3D environment (Section 4.3). The system
architecture, which combines these elements to create newer and newer paths
through the repository, is described in Section 5. Discussion and conclusion end
the paper.
1https://www.google.com/culturalinstitute/u/0/project/art-project
2http://app.labyrinth-project.it:8080/LabyrinthTest/
2
2. Background
In the last decade, the use of ontologies for the access to cultural heritage
collections has been investigated by several projects. A pioneering contribution
was given by the Finnish Culture Sampo project [16]. In this project, a number
of domain ontologies provide the background against which cultural objects (in-
cluding artworks, artists, traditional practices, etc.), encoded in di↵erent media
formats (e.g., images and videos), can be explored, tracking the underlying rela-
tions over them. In CultureSampo, once a certain artifact (e.g., a painting) has
been retrieved, it is possible to explore the relations over the objects (and char-
acters) represented therein. The system has recently evolved towards a linked
data approach with the release of a new application, War Sampo, focused on
the Second World War [9]. The Agora system [17] frames the exploration of
a digital collection into historically relevant episodes, supported by a semantic
account of the notion of event [18]. For example, the user can choose a historical
episode (e.g., “German occupation of Poland in the Second World War”) and
navigate among the cultural objects related to this event.
A line of research in ontology-based systems has specifically explored the use
of narrative models in cultural heritage dissemination. Stories not only repre-
sent an e↵ective way to convey information in a compact format, as argued by
[19], but, according to the research in cognitive psychology, they are a primary
means for the conceptualization of reality [20]. In cultural heritage, many art-
works have, by and large, some type of narrative content. In visual arts, for
example, paintings often display story episodes while statues immortalize char-
acters; even non representational artworks often refer to narrative elements,
despite the abstract nature of their visual content. Stories are narrated by tex-
tual media such as tales and novels, but also – though in nonverbal terms – by
di↵erent kinds of musical works, from operas to symphonic poems. Narrative
is the focus of the Bletchley Park Text system [21], a semantic system designed
with the goal of supporting the users in the exploration of online museum col-
lections. Designed with the notion of the “guided visit” in mind, the system
encompasses an ontology of story, taken from the Story Fountain project [22].
The stories represented in the system are employed in a web interface to create
relations over entities in online collections; based on this knowledge, the user can
ask the system to find a narrative connection between di↵erent entities. More
recently, the Decypher EU project leverages stories to addresses the curatorial
side of cultural heritage dissemination [23]. In Decipher, a story ontology is
the basis of a system that supports the creation of story-based collections by
museum curators. Finally, Europeana also uses some simple narrative features
to describe the items they contain. In Europeana, it is possible, for example,
to navigate among the artifacts representing a given action or displaying a cer-
tain character, across a large number of indexed objects; the system does not
provide, however, a story–level navigation.
The Labyrinth project extends the approaches described above by integrat-
ing the use of a narrative model to connect the items in a collection with the use
of a visual environment for the exploration of these connections. The system re-
3
lies on an ontology of narrative archetypes to describe the items in the collection;
the 3D interface of system is inspired to a well known narrative archetype, the
“labyrinth”. The notion of labyrinth is not only deeply rooted in the Western
Culture, dating back to Greek Myths and witnessed by several archaeological
locations across Europe [24], but also, thanks to the graph like nature of the
notion of labyrinth [25], it lends itself well to representing the many-to-many
relations among artworks encoded in the ontology. The goal of the visual design
of the 3D interface is two-fold: on the one side, it is aimed at engaging the users
to explore the repository though an immersive experience; on the other side, it is
aimed at making the system usable by the large majority of uses by integrating
information giving and entertainment in a familiar environment. The labyrinth,
or maze, is a genre of video games most users are familiar with, thanks to classic
2D games such as Atari’s Pacman3 and recent 3D titles such as Imangi’s Temple
Run4 or PlayFirst’s Dream Chronicles5.
In cultural heritage, the use of 3D visualization is normally intended as a
support for study and dissemination activities. 3D projects in cultural heritage
can be roughly divided into two types: virtual equivalents of physically exist-
ing locations, such as museums and historical buildings, and reconstructions of
physical environments that have disappeared, such as archaeological locations or
temporary art works. Google Art Project6 and Arounder7 are examples of the
first type, where 3D is often obtained through PMVR techniques that integrate
high definition images of artworks in the 3D environments. Rome Reborn [26],
the 3D reconstruction of Rome as it appeared in the IV century, is an example
of the second type. In this project, the use of 3D is integrated with animated
characters of ancient Romans, who interact with the users. A similar approach
is proposed by [27], who present a framework for 3D real time applications in
web browsers, employed to develop virtual reconstructions of Rome (Virtual
Rome project) and other Italian locations [28]. Labyrinth di↵erentiates from
these approaches since the 3D representation is not employed to reconstruct
real environments or to create virtual ones, but as a tool to convey semantic
relations through a visual environment. For this reason, the system does not
encompass a semantic model of the 3D environment: rather, it maps a semantic
representation of the domain onto the 3D environment, as part of the interaction
design process.
3. Live System
The Labyrinth project encompasses both a standard web based interface [29]
and a 3D application [30]. Both interfaces allow the user to navigate a repository
of cultural objects with the guidance of a set of archetypes of narrative nature.
3https://en.wikipedia.org/wiki/Pac-Man
4http://www.imangistudios.com
5http://www.playfirst.com/games/view/dream-chronicles
6www.google.com/culturalinstitute/project/art-project
7www.arounder.com
4
Figure 1: A screenshot of the web interface of the system (in Italian).
The archetypes are contained in an ontology that describes each archetype in
terms of its related stories, characters, objects, events, and locations, and stores
the connections that relate these categories with the items in the repository.
In both the hypertextual and the 3D interfaces, the interaction with the
user starts with the selection of an archetype. In the hypertextual interface,
the users continues by refining her/his search based on the inner articulation of
the selected archetype into more specific categories (namely, stories, characters,
objects, events, locations and epochs), then into single elements within the
category (single story, character, etc.), by following a top-down strategy that
ends with the selection of a specific artifact (for a detailed description, see
[31]). Fig. 1 shows a screenshot of the interface (in Italian): after selecting
the archetype of the labyrinth, the user has decided to explore the category of
“stories”, then the specific story entitled “Theseus kills the Minotaur”. As a
result, the interface shows a record of the story (upper part of the main box,
“Teseo uccide il Minotauro”), which includes the related stories, the characters
and objects featured by the story, and the locations and epochs in which the
story takes place. The user can click on them to navigate from the currently
selected story to another, or to move to a di↵erent category of the archetype
(for example, “characters” or “locations”). Below (bottom of Figure 1), the
interface shows the thumbnails of the artifacts (or, better, of their digital copy)
that refer to the currently selected element (here, the story “Theseus kills the
Minotaur”); each thumbnail can be clicked on to get a record of the artifact.
5
Figure 2: 3D interface: selection of the archetype (left) and assignment of initial and target
artworks (right, labeled as “current” and “target”).
Figure 3: First step of the navigation: the initial artwork, Minotauromachia (left); right: some
of the doors available from the initial artwork (same character and same story).
A slide show of the thumbnails is positioned on the left of the story record to
provide a quick glance on the available contents for the current selection.
Di↵erently from the hypertextual interface, the 3D interface is characterized
by a bottom-up approach: here, the user navigates from artifact to artifact on
the basis of the relations over them represented in the ontology, building her/his
own personal path through the repository. The user is situated in a virtual maze
where the artworks are located in the clearings and connected by pathways that
represent the relations over the artworks. Immersed in the virtual maze in a
first person perspective, the user is encouraged to explore the repository in the
same way as the visitor of a hedge maze explores the turns and twists of the
maze on her/his way to the exit.
In order to exemplify the user experience in the 3D labyrinth, we will describe
a navigation example extracted from the system log, illustrated through the
screenshots of the steps that compose it (Figures 4, 5, 6). After choosing the
archetype of the “labyrinth” (Fig. 2, left), the user is assigned a start and
a target artwork (Fig. 2, right), randomly extracted from the repository: in
the example, they are, respectively, the “Minotauromachia” (a painting) and a
novel, “Il labirinto greco”. When the user clicks on the Start button posited
below the start and target nodes, she/he is brought to the 3D environment
6
Figure 4: Second step of the navigation: a pathway (left) to the subsequent artwork; second
artwork (right), with doors leading to the other artworks in the “same character” relation.
Figure 5: Third step of the navigation: a Greek vase representing Thesues killing the Minotaur
(left); doors leading to the artworks referring to the same story (right).
(Fig. 3, left). The first location is the node containing the start artwork, the
painting entitled “Minotauromachia” by Pablo Picasso, which shows the Greek
hero Theseus fighting with a Minotaur. Fig. 3 (left) shows how the artwork
(here, a picture of the painting) is displayed to the user in a 2D layer temporarily
superimposed to the 3D scene; the artwork is accompanied by the information
about its author, the place where it is hosted and the creation date. A longer
description can be obtained by clicking on “description”, below the image; by
clicking on “close”, the layer disappears. Figure 3 (right) shows some of the
connections available from the node, represented by the two doors labeled as
“agent” and “story” (other doors are out of the view): the first door leads to
artworks that feature the same character (named “agent” in the system) as the
current artwork, the second door leads to artworks that relate to the same story.
By choosing the door labeled as “agent”, the user is led through a path-
way (Fig. 4, left) to an empty node (Fig. 4, right) that contains doors for
the artworks that display the same character as the previous artwork, The-
seus; the titles of the artworks are written above the doors, from left to right:
“Monete rivenute a Cnosso” (“Coins found in Knossos”), “Teseo uccide il Mino-
tauro” (“Theseus kills the Minotaur”), “A↵reschi della villa imperiale a Pompei”
(“Frescos, Villa Imperiale in Pompei”). The user selects the middle door, and
7
Figure 6: Left: fourth step of the navigation, a statue of Ariadne (same story relation); a
di↵erent artwork (right), reachable by backtracking to the previous step (a Roman fresco
representing the myth of the Minotaur in Pompei).
is led to a node that contains a Greek vase displaying the Theseus in the act of
killing the Minotaur (Fig. 5, left). Notice that the console posited in the bottom
part of the interface contains, besides the controls for getting help, stopping the
sound and exiting the application, a progress bar displaying the artworks visited
by the user so far: by selecting a previously visited artwork, the user is brought
back to the node containing it. After the Greek vase, the user follows the same
story relation by clicking on the door labeled as “story” (not shown), and is
brought to an empty node with doors for the artworks that refer to the same
story: “Arianna dormiente” (“Sleeping Ariadne”) and “A↵reschi della villa im-
periale a Pompei” (“Frescos, Villa Imperiale in Pompei”: notice that, like in a
true labyrinth, the same artwork can be gained by following di↵erent paths).
By choosing the first door, “Arianna dormiente” (“Sleeping Ariadne”), the user
will reach a node containing a Roman statue of Ariadne, the female character
of the myth of the Minotaur (Fig. 6, left); from there, by backtracking to the
previous node, the user may select the second door, “A↵reschi della villa impe-
riale a Pompei” (“Frescos, Villa Imperiale in Pompei”), which leads to a node
containing a painting that illustrates the myth of the Minotaur, located in a
Roman villa in archaeological site of Pompei (Fig. 6, right).
4. The tripartite core of Labyrinth
Given a collection of cultural objects, commonly represented by the digital
resources that constitute the “digital equivalents” of the actual physical objects
[32], the access to the collection in an ontology based system such as Labyrinth
3D is the result of the interplay of three elements, namely the information about
the objects, or metadata, contained in the ontology (Section 4.1), by which
the objects are indexed, the visualization interface (Section 4.2), driven by the
project specific goals (dissemination, presentation, study, etc.), and the mapping
of the objects onto the visualization interface (Section 4.3). Thanks to this
tripartite relation, the system translates the information about the cultural
objects into a visual representation where the semantic relations over the objects
contained in the ontology are mapped onto the elements of a 3D environment.
8
4.1. The Archetype Ontology
The description of the artworks encoded in the metadata typically includes
features such as date, authorship and title of the items, normally expressed ac-
cording to standard vocabularies, such as ISO 8601 for dates8 or ULAN (Union
List of Artists Names) for names.9 Beside authorship and editorial informa-
tion, metadata usually contain also information about the management and
preservation of cultural objects, such as responsibility for the preservation, dig-
italization standards, etc. In the last decade, metadata have evolved towards
semantic encodings describing the content of the artworks, with categories such
as iconography, event types, etc.[33, 34]. An example of semantic description is
provided by the Europeana Semantic Model (ESM)10.
As exemplified by the navigation example provided in the previous section,
the semantic annotation of the artworks in Labyrinth is mainly oriented to the
representation of their content, and narrative content in particular. The narra-
tive content of the artworks is expressed through a set of archetypes that char-
acterize Western culture through the ages, heritage of the Greek and Roman
tradition [35]. The core of the Labyrinth system is the Archetype Ontology
(AO), described in detail in [31]. The AO contains a number of archetypes
(the journey, the labirinth, the hero) and describes how the artworks relate to
them via the representation of stories, characters, objects, events, locations and
epochs. The AO contains 8 top classes: the Archetype class contains the
archetypes; the Artifact class contains the artworks, organized according to
the FRBR model [36]; Entity contains the characters and objects represented
in artifacts; Story represents a collection of stories related with the archetypes;
the Description Templates class contains a role-based schema for describ-
ing events and states that can be filled by characters and objects; the Format
class encodes the format and type of media resources; Geographical Place
and Temporal Collocation, finally, encode, respectively, the spatial and tem-
poral information related to artifacts, stories and archetypes. The Archetype
Ontology was manually built based on an extensive survey of the notion of
archetype, spanning from Warburg’s Bilder Atlas [14] to folkloric studies [15]
and contemporary accounts of tropes in media.11 The ontology was aligned
with the conceptual reference model established by the International Council of
Museums (ICOM), the CRM-FRBR model [36], a standard in the description
of cultural heritage.
In the Labyrinth system, the editing phase is conducted through a back-end
web interface through which items can be added to the repository. The descrip-
tion of the items is accomplished through form filling and it follows the Dublin
Core initiative [37], a metadata element schema that has become a standard de
8http://www.iso.org/iso/home/standards/iso8601.htm
9http://www.getty.edu/research/tools/vocabularies/ulan/
10http://pro.europeana.eu/ese-documentation
11http://tvtropes.org
9
facto in digital archives12. When a new item is added to the repository, the
system imports the description of the item in the ontology through a built–in
procedure that converts the input data into the ontology format (a set of RDF13
triples). First, the internalization phase (described in details in [29]) translates
the metadata of the resource (creator, date, etc. ) into the language of the
ontology. Then, a mapping procedure matches the imported description with
the available archetypes. Both steps are achieved via if-then rules encoded in
SWRL14, the rule language designed for ontologies. As an example of how the
mapping is accomplished, consider the rule that examines the “title” metadata
element of an artwork in order to find a connection with the archetype of the
“Labyrinth”: if words like “labyrinth” or “maze” are found in the title, the rule
will add to the ontology the assertion that the artwork evokes the archetype of
the labyrinth. Finally, after the new item has been internalized in the system
and mapped onto the ontology, a specific set of rules add the narrative features
to the artworks (narrative mapping). For example, if an artwork represents a set
of characters performing some action, the system searches for a story in which
the same characters perform that action (see [29] for a detailed description of
how the narrative properties are added to the representation of the items). For
instance, an artwork representing Ariadne in the act of giving the ball of thread
to Theseus (a focal event in the myth of the Minotaur) would be recognized as
having the myth of the Minotaur as a narrative component.
In order to illustrate how the items in the repository are represented in the
ontology as a result of the internalization process, we will resort to an exam-
ple. Fig. 7 illustrates the description of the painting “Minotauromachia” by
Pablo Picasso (the first step of the navigation example in Section 4.1), serial-
ized in the XML/RDF format. Notice that each line represents an RDF triple,
composed of a subject (all triples in this fragment have the same subject, i.e.,
the named individual “Minotauromachia” in line 2), a predicate (for example,
hasResourceType, line 4) describing a property of the subject or a relation with
another entity, and an object (here, the resource type, Image) which constitutes
the value of the property – or the second term of the relation. All resources are
characterized by a prefix given by the URI of the ontology. Fig. 7 outlines the
role of each phase of the procedure described above. The annotation is divided
into two groups of assertions: the first group contains the assertions extracted
from the artwork metadata by the internalization procedure; the second group
contains the properties added by the mapping procedure, which connects the
artwork with the archetypes: this group contains specific annotations concerning
the narrative relations among the artworks.
Lines 3 to 6 are created by the internalization procedure and correspond to
the artwork metadata, such as type, creator, etc. The hasResourceType prop-
erty (line 4) describes the media type of the resource, i.e., image; the hasCreator
12https://www.iso.org/obp/ui/#iso:std:iso:15836:ed-2:v1:en
13https://www.w3.org/TR/rdf-concepts/
14https://www.w3.org/Submission/SWRL/
10
Properties added by the internalization phase:
1
2
3
4
5
6
After the mapping phase:
7
8
10
After the narrative reasoning:
11
12
Figure 7: The description of the artwork “Minotauromachia” by Pablo Picasso in the AO
ontology. The sections show the properties added by each phase of the internalization and
mapping procedures.
property (line 5) connects the painting with its author, “Pablo Picasso”; the
hasGeographicalLocation property (line 6) describes the location of the artwork.
Lines 7 to 11 describe the relation of the artwork with the archetype: the prop-
erty evokes (line 7) relates the painting with the archetype of the “Labyrinth”,
while a set of specific properties describe the relation with the archetype in
greater detail, focusing on its narrative aspects: displays (lines 8-9) refers to the
characters which appear in it, i.e., Theseus and the Minotaur; describesAction
(line 10) refers to the event type it depicts (“killing”). Finally, the property has-
Part (line 11) states that the painting contains, as part of its narrative content,
the Minotaur Story.
Given this description, several relations can be detected with other artworks.
Besides the standard relations based on author or resource type, the archetype
of the labyrinth connects the artwork with other artworks that display the same
characters (Theseus or the Minotaur), depict the same action type (killing), or
refer to the same story (the myth of the Minotaur) and other related stories
(e.g., Ariadne and the Thread).
4.2. Designing the 3D environment
The design of the 3D environment is inspired by the metaphor of the labyrinth.
This metaphor was chosen for its ability to convey the graph–like nature of the
relations over the artworks in a cultural heritage collection, and for its imme-
diacy of use, since it provides an intuitive mapping for artworks (nodes of the
11
labyrinth) and relations over them (connections among the nodes). The interac-
tion metaphor underlying the navigation is “finding one’s way”: here, however,
the user does not simply gain the exit, but the creation of a personal path in
artworks’ meaning, represented by a virtual “red thread”. In order to make the
experience more engaging, when the session begins, the user is given a target
node. When the user reaches the target node, or when the user decides to exit
from the labyrinth, the session ends and the user is shown the statistics about
her/his own path: number of visited nodes, elapsed time, backtrackings, etc.
The visual design of the labyrinth is inspired to the classical hedge maze,
with architectural elements that are intended to remind of some distant but
indefinite past; this choice was primarily due to the constraint posed by the
heterogeneity of the contents assumed by the project. The floor is partly tiled,
partly covered with grass, and the mood is inspired by a dark, Gothic style. The
maze contains two types of nodes: some nodes (artwork nodes) contain artworks,
some nodes (relation nodes) are empty and only serve the function of connecting
the artwork nodes, as exemplified in the navigation example provided in Section
3. The presence of an artwork in a node is signaled by a low circular balustrade
in the middle, open in several points, that are intended as a↵ordances inviting
the user to step into the inner part of the node [38]. The entrance to pathways
is marked by doors; each door corresponds to a semantic connection (e.g., same
story), and is surmounted by the name of the type of connection (e.g., “story”).
Each node has a fixed number of doors/pathways: depending of the number
of semantic relations that connect the node with other nodes, some doors may
be closed, or hidden by greenery. If the connection leads straight to a single
artwork, the title of the artwork is posited above the door. Pathways di↵er in
length and form: some are short, some are longer and they bend, so that their
end is not visible, in order add some thrill to the experience.
The navigation in the system is inspired to the paradigm of constrained
navigation [39], with the aim of making it usable also for non expert users of
3D applications. The user moves by clicking on small circles of light posited on
the floor, in front of the doors of the nodes and along the pathways. Circles of
the same type also mark the presence of an artwork in the middle of artwork
nodes and must be clicked to get information about the artwork. Smaller circles
of light appear inside the circles when they are clicked, so that they eventually
form a sort of “red thread” that marks the path made by the user so far. The
metaphor of the red thread, aimed at improving self orientation, is enforced also
by the console posited in the lower part of the screen, that shows the list the
nodes visited by the user. By clicking on a node in the list, the user is brought
back to that node. The console also contains buttons for ending the session and
turning o↵ the sound. The user is free to explore the labyrinth, going back to
previous locations and clicking on the control posited in artwork nodes to receive
information and experience them via the appropriate plugins: depending of the
media type of the resource associated with the artwork, an image is displayed,
a video is played, etc. A short description of the artwork, with title, date and
creator, is always provided, as exemplified in Fig. 3 (Sect. 3).
12
The 3D environment was implemented with the Unity 3D15 real time engine,
which supports several platforms and mobile conditions. Unity 3D o↵ers first
person gameplay default assets, both concerning camera motion control and
mouse tracking motion control. In order to optimize the production time and
cost of the 3D assets, a single model of the labyrinth node, with a predefined
set of exits, was created: at run time, it is dynamically adapted to the semantic
relations connecting the current artwork with the others by closing or opening
the corresponding number of doors. To achieve our goals, we built an indexed
database of 3D objects to be promptly displayed in real time by the 3D Engine.
Thus, we were able to produce several theme variations exponentially explod-
ing the number of possible combinations. The standard 3D objects are: the
octagonal square (3 variants, actually), the open door (3 variants), the closed
door (2 variants), the textual artwork viewer (1 variant), the pictorial artwork
viewer (1 variant), the movie viewer (1 variant). The pathways are a 3D ob-
ject category on their own: they vary in shape in accordance to their length,
which is measured in steps (2-3-4 steps, each in three variants). Joined together,
steps compose asymmetrical paths, which can also be used backwards, therefore
multiplying the possible combinations of subsequent pathways. So, the maze,
determined through the user choices, is perceived as never being the same.
4.3. Mapping the ontology onto the 3D
The mapping of the semantic relations onto the visual environment poses
some problems that need to be faced as part of the system design, and constrain
the architecture of the system. Formally, the labyrinth is an undirected graph
[25], where vertices have a variable degree.16 The nodes correspond to the
graph vertices, the pathways to the edges. Notice that, as in a real maze,
there are also nodes with only one incident edge, i.e. dead ends where the user
has to backtrack. The direct transposition of the graph-like structure of the
relationships over the artworks from the ontology to the 3D labyrinth, however,
would lead to a proliferation of the edges that would be confusing for the user.
Take, for example, the similarity relation “displaying the same character” among
artworks. Representing this relation as artwork to artwork relations implies that,
for each artwork, an edge should be added from the artwork to every other
artwork that displays the same character (and this should be done for each
semantic relation). In order to alleviate this problem, in Labyrinth, we decided
to represent the semantic relations such as “displaying the same character”
through special nodes that represent the relation itself, thus obtaining a more
compact representation. These nodes do not correspond to artworks, but simply
distribute the semantic relation over the pairs of artworks that are in the given
relation.
15https://unity3d.com
16For usability reasons, the maximum number of edges per vertex has been limited to the
arbitrary threshold of 7, given the well known limitations of short term memory first shown
by Miller [40].
13
Reasoner
Triple store
Archetype
ontology
Media
storage
Resource
descrip6on
3D Labyrinth
Web interface
Seman6c naviga6on
Ontology-to-3D
API
ONTOLOGY SERVER
WEB SERVICES
VISUALIZATION CLIENTS
DC
RDF/
XML
Client side
Server side
DB
SWRL RULES
mapping
internaliza6on
MEDIA REPOSITORY
Figure 8: The architecture of the Labyrinth system.
As a result of the constraint described above, there are two types of nodes
(with di↵erent iconic elements) in the labyrinth, connected by the pathways:
artwork nodes and relation nodes. Artwork nodes are connected with both
relation nodes and artwork nodes. Relation nodes are connected only with
artwork nodes. The user navigation starts from an artwork node: the user
has to choose one of the pathways exiting from the node, labeled either with
the name of a di↵erent artwork (in this case, the pathway, leads directly to an
artwork node) or with the name of a semantic relation (in this case, the pathway
leads to a relation node, that in turn leads to a set of di↵erent artwork nodes).
Since the semantic relations are symmetric, pathways can be walked both ways.
5. The Labyrinth system
In this section, we describe the architecture of the labyrinth system, which
constructs the 3D environment as long as the interaction with the user pro-
gresses.
5.1. System architecture
The architecture of the Labyrinth system is structured according a client–
server schema. The ontology is stored in an ontology server; the information it
14
contains is dynamically extracted from the ontology and made available to the
visualization client, which manages the user interface.
The system encompasses four main modules (see Fig. 8):
• the Ontology Server (Fig. 8, top) stores the AO ontology – where the
cultural heritage objects are described – and provides the reasoning ser-
vices that allow the system to establish the relations of each object with
the archetypes, as exemplified in Section 4.1 (for example, inferring the
relation between an artifact and a story given the characters displayed in
the artifact). In the current implementation, the ontology server is pro-
vided by Owlim.17 The ontology server also supports the SWRL rule sets
that implement the internalization and mapping procedures described in
Section 4.1, by which new items are ingested in the system. The ontology
server provides also the SPARQL18 endpoint for querying the ontology,
necessary to extract from the ontology the data that will be visualized in
the interface (i.e., the semantic relations over the artifacts, such as the
same character or the same story relations exemplified in the navigation
example in Sect. 3). Notice that this module is independent of the vi-
sualization type and it serves both the web-based interface and the 3D
environment.
• the Media Repository (Fig. 8, right) stores the media objects (the
digital equivalents of the artworks) which constitute the repository of the
system and is indexed by a relational database (a mySql database);
• a set of Web Services (Fig. 8, left) implement the Application Pro-
gramming Interface specific to each visualization client. This component
extracts the data from the ontology in response to the requests of the
clients. The web services, written in Java, are called by the visualization
clients to respond to the actions of the user, and return the data in XML
format. For example, in the 3D environment, when the user clicks on a
door leading to a given artwork, the visualization client calls the API com-
mand that fetches the information about the artwork, needed to generate
the node with the artwork in the 3D maze.
• the Visualization Clients (Fig. 8, bottom) support the interaction with
the user through 3D navigation, as standalone application (for the 3D
system) or embedded in a web browser (for the web interface).
The core of the system consists of the APIs that fetch the data from the
ontology to the visualization clients. The interplay among these components
realizes the mapping of the objects and relations encoded in the ontology onto
the environments where they can be visualized by the user. In the following, we
describe in detail the interaction between the Ontology-to-3D API and the 3D
17http://www.ontotext.com/products/ontotext-graphdb-owlim-new-2/
18https://www.w3.org/TR/sparql11-overview/
15
visualization client. The interaction between these two components achieves the
computational creativity that the user can enjoy by navigating the repository
in the virtual maze.
5.2. The system at work
The topology of the maze is computed locally as the user progresses in
her/his path. This choice is partly related with the user experience design and
partly related with optimization issues. Concerning the user experience, the
step by step generation of the maze provides room for the adaptive personal-
ization of the navigation experience, which can be tailored to the typology and
behavior of the user given the available relations over the artworks. Currently,
the variability of the navigation is provided by a basic random mechanism. Since
the system does not pose any constraints on the number of related artworks, the
available relations over them may exceed the number of doors, set to 6 in the
current implementation, when necessary, 6 artworks are randomly extracted. As
a result, in di↵erent navigation sessions, di↵erent artworks may be extracted,
thus generating slight variations in the user experience.19 Concerning the op-
timization issues, the step by step generation of the topology guarantees that
the computation needed to generate the maze is not a↵ected by the size and
the connectivity degree of the repository (which only a↵ect the execution of
the queries).20 Moreover, if the repository changes, no initialization is needed.
Notice also that this solution is made possible by the fact that the 3D envi-
ronment does not encompass a top down, map-like visualization of the maze,
with the consequence that the user can only experience a subjective view of the
environment. This choice, although debatable for the lack of orientation it may
provoke in the user, is consistent with the actual experience of the real hedge
maze by which the design of the 3D system is inspired.
Basically, the maze is generated on the fly as follows: when the navigation
begins, the system retrieves from the ontology the information about the first
artwork, and generates only the portion of the maze which describes this artwork
and its connections with the others, namely a node containing the artwork
and the pathways which represent its semantic relations with the others. Each
pathway represents a relation type (e.g., story, location or agent/character),
as illustrated by Figure 3 (right), where the doors leading to the pathways
for “story” and “agent” are visible. When the user makes the next choice by
selecting the pathway she/he wants to take, the next portion of the maze is
created. If the selected pathway leads to a group of artworks (i.e., the relation
it represents contains multiple artworks), the system creates an empty node
whose function is to redirect the user to the single artworks, each placed in
a di↵erent node (see the example in Fig. 4, right, with doors for the single
19Notice that, if the user backtracks in the same session, the nodes that have been already
visited are not generated from scratch, to let the user orientate her/himself.
20The prototype ontology currently contains 1211 triples, but a wholly functioning system
would be much larger.
16
Figure 9: The interplay between the ontology server (left) and the interaction with the user
(right) operated by the Ontology-to-3D API.
artworks). A direct pathway is generated only if the selected pathway leads
to a single artwork. The rationale behind this strategy is to enforce the 1 : 1
mapping between artworks and nodes, so that a node always contains a single
artwork.
In the following, we describe the algorithm executed by the 3D visualization
module to manage the interaction with the user, supported by the ontology-
to-3D API (see Figure 9). The visualization client (3D Labyrinth, bottom of
figure) queries the ontology (top of the figure) through the ontology-to-3D API
(see Section 5.1). The generation of the 3D labyrinth is accomplished through
the following steps:
Initialization. The session begins when the user starts the application on the
client device.
• Session start. First, the 3D client queries the ontology server to get the
list of the available archetypes through the startLabyrinth3D() command
(Fig. 2, left).
• Archetype selection. When user chooses one of the available archetypes,
the client sends the selected archetype to the server (setArchetype()).
17
• Navigation initialization. The client invokes an initialization command
(initialize()) that selects a random pair of artworks: they provide, respec-
tively, the initial and target nodes (Fig. 2, right). When the user clicks
on the “start” command in the interface, the navigation begins.
Node generation At this point, the next node is generated, until either the
user reaches the target node or she/he exits the labyrinth (by clicking the exit
button posited in the navigation console). This loop is repeated each time the
user selects the next artwork.
1. Retrieval of relations. The client queries the ontology to get the informa-
tion about the chosen artwork (or the initial artwork at the beginning of
the navigation) through the command getNodeInfo(). This command is
the key to the mapping of the semantic relations encoded in the ontology
onto the 3D labyrinth: given an artwork, it returns the identifier of the
digital resource that represents the artwork in the media repository, its
metadata (the information about its creator, title, etc.) and the list of its
semantic relations (character-based, story-based, location-based relations,
and so on) with the other artworks. The client employs the information
about the artwork’s relations with the other artworks to build the node
that will contain the artwork and stores the digital resource and the infor-
mation about the artwork to generate the 2D panel describing the artwork
in case the user requires it (as exemplified in Fig. 3, left, Fig. 5, left, and
Fig. 6).
To retrieve these data from the ontology, getNodeInfo() executes a set of
SPARQL queries on the ontology, one for each possible type of semantic
relations. For example, the following query extracts from the ontology
the set of artworks ?a that are evocative of the archetype of the labyrinth
(: evokes) and display the character of Theseus (displays):
SELECT ?a
WHERE{
?a :evokes :Labyrinth
?a :displays :Theseus
}
By executing similar queries for all the semantic relations embedded in the
system (namely, story, character, event, location, epoch and object), the
system collects all the available relations connecting the selected artwork
with the rest of the repository.
2. Computation of topology. The method getNodeInfo() returns an XML
fragment describing the selected artwork and its related artworks; in prac-
tice, the XML contains a section for each semantic relation type (agent,
story, etc.). For example, consider the following fragment, returned by
invoking the command getNodeInfo() on the artifact displayed in Fig. 5
(left), a Greek vase of the 5th century b.c. displaying Theseus killing the
Minotaur:
18
Minotauromachia_Picasso
...
Ariadne_and_the_Thread
Sleeping_Ariadne
...
VillaImperiale_Pompei
The response contains the set of artworks (or artifacts, as artworks are
generically termed in AO) related to the input artwork, indexed by re-
lation types: the example shows the story relation (artworks tagged as
rartifactstory) and character relation (artworks tagged as rartifactagent).
The example response contains (among other artifacts not listed in the ex-
ample) two story-related paintings, “Minotauromachia” by Pablo Picasso
and “Ariadne and the Thread” by the Italian painter Palagio Pelagi, and
two character-related artworks, namely an anonymous statue representing
a sleeping Ariadne (situated at the Vatican Museums) and the frescoes
depicting the myth of the Minotaur situated at the “Villa Imperiale” in
Pompei. Notice that the latter two artworks are displayed in Fig 5 (right)
as available alternatives after the user has chosen the story relation from
the node containing the Greek vase (Fig 5, left): in the figure, they are
termed, respectively “Arianna dormiente” (“Sleeping Ariadne”) and “Af-
freschi della Villa Imperiale” (“Frescoes, Villa Imperiale”).
At this point, the topology of the labyrinth is computed. For each semantic
relation (in the example response: and ):
– if the relation contains a single artifact, an artwork node is created to
represent it and a pathway is added from the chosen node to the new
artwork node;
– if the relation contains multiple artifacts (as in the standard case), a
relation node is created and a pathway is added from the chosen node to
the relation node; for each artifact, then, an artwork node is created and a
pathway is added from the relation node to each of the new artwork nodes
(see Fig. 3, right).
3. Generation of the labyrinth. Based on the topology computed above, the
next node of the 3D labyrinth is created and added to the 3D environment,
19
together with its exiting pathways. When the user chooses a new artwork
(either directly connected to the current one or indirectly, via a relation
node), the loop is repeated.
End of session. When the user either reaches the target node or clicks on the
exit button, the client executes the endLabyrinth() command to visualize the
statistics of the session (time elapsed, visited nodes, etc.) and closes the session.
6. Lesson learned
We carried out an evaluation of the 3D interface of the system, in order to
gather information about the users’ liking of the system and their expectations
about its use. The evaluation took place in a scientific fair, with some users
taking part to participated demos and some other user freely interacting with
the system. The experimentation is described in detail [41]: here, we only
summarize the most important results, which are relevant for discussing the
potential and the possible applications of Labyrinth 3D. 41 testers took part in
the evaluation, males and females, with ages ranging from 10 to 67 years old.
The system was very well welcomed by the visitors of the fair, in particular by
students and teachers, who were enthusiastic about its potential for education
and dissemination. The ethnographic observation of the testers who interacted
directly with the system showed that the navigation was generally easy, with
some problems in clicking the navigation controls when they were located far
away along the pathways, because the distant controls tend to be small due to
the perspective. Users were sometimes bewildered at finding themselves in a
node they had already visited, but were ready to accept the explanation that
this is typical of labyrinths. The users tended to read carefully the information
displayed about the single items, reasoning aloud about their connection with
the archetype and with the previously visited nodes. A questionnaire was given
to the users to assess their liking of the system and their preferences about
the use contexts. The questions about the use of the system revealed that
the users would prefer the PC and the tablet for using the system, a finding
that is in line with the goal of the project of creating an immersive experience.
When asked about the similar media, the users selected the video game and
the encyclopedia, also in line with the design goal of creating a tool for cultural
dissemination. In particular, a group of 6 questions were aimed at investigating
the general acceptance of the system: by using Likert scales (with 5 points
from �2 to +2, mapped onto values from 1 (�2) to 5 (+2) in the subsequent
data analysis), we asked testers to what degree the system was: i.intuitive,
ii.interesting, iii.engaging, iv.useful, v.appealing, vi. straightforward to use. The
average value of the answers to the questions concerning the acceptance was 4.5,
with “interesting” as the highest average value (4.7) and “straightforward” as
the lowest average value (4.32), indicating that the application was appealing
but that its use was not entirely clear to some users. The values are illustrated
in Table 1. As it can be noticed, the standard deviation is not high, meaning
that the testers generally agreed on a positive evaluation.
20
Table 1: Average values for the questions about perceived properties of the system, on a 5
point Likert scale.
subquestion SYSTEM PROPERTY AVERAGE VALUE ST. DEV.
i intuitive 4.35 0.72
ii interesting 4.7 0.57
iii engaging 4, 41 0.74
iv useful 4.48 0.61
v appealing 4.5 0.74
vi straightforward 4.32 0.68
The results of the evaluation suggest that the proposed approach works and
open the way to a re-use of the architecture of the system for applications that
leverage the creativity intrinsic to a cultural heritage archive to generate per-
sonalized paths in a 3D environment. A precondition to the reuse the approach
of the Labyrinth project is to abstract the experience in the design and imple-
mentation of the system into a pipeline for creating similar applications. Given
our experience in the design and implementation of Labyrinth 3D, we propose
the following pipeline, divided into three phases: visual design, software de-
velopment and editing, each characterized by specific professional roles. We
skip the conceptual modeling phase, assuming that an annotated repository of
cultural heritage objects is already available (for example, as part of some an-
notation project or as a by product of a digitalization initiative). It is possible,
however, that, for specific projects, an ad hoc ontology is developed to satisfy
this requirement: if this is the case, an ontology engineer and a domain ex-
pert cooperate to design the ontology that will constitute the backbone of the
system.
The visual design phase is aimed at bridging the gap between the concep-
tualization of heritage objects (the archetypes in Labyrinth 3D) and the users
through the use of visual and spatial metaphors (the maze in Labyrinth 3D).
As argued by [13], the choice of the metaphor is crucial to communicating the
conceptual model. This phase should be conducted with the help of a sample
repository where a few objects have been inserted to support the design process
and the subsequent development phase. Given the annotated repository, the
interaction designer, in cooperation with a visual designer, i) devises a suitable
metaphor for conveying the description of the objects in the repository through
the 3D environment (by mapping of the object properties and their relations
onto the features of the environment), ii) designs the interaction flow (speci-
fying how the user can interact with the 3D environment and what responses
he/she should get in each phase of the interaction) and iii) establishes the visual
properties of the 3D, such as its mood and appearance. A game designer may
be involved in this phase to insert elements of playability into the interaction.
As shown by [42], in fact, the use of game in tandem with visual metaphors
increases the levels of learning.
21
The software development phase translates the interaction design into 3D
assets, staged and manipulated by a 3D engine. Once the interaction metaphor
has been established, the 3D models that constitute the environment are cre-
ated and arranged in a set of layouts by a 3D production team, together with
animations and camera movements (in case the navigation is achieved by con-
straining the user to predefined movements, as in Labyrinth 3D). In parallel,
the semantic web developer implements the queries that extract the object de-
scriptions from the ontology (previously uploaded onto an ontology server) and
makes them available by programming a web service available though an API.
Finally, the 3D developer programs the 3D environment so that it implements
the interaction flow established in the interaction design phase.
In the editing phase, the cultural heritage objects are collected and anno-
tated with the semantic metadata required by the conceptual model encoded
in the ontology before adding them to the repository. Metadata may include,
for example, the relations of heritage objects with locations, artists, historical
events, etc. Although professional annotators are preferred, metadata may be
also contributed by amateurs through crowd sourcing, as recently proposed by
[43].
7. Conclusion
In this paper, we described Labyrinth 3D, a system where the user can
explore the semantic relations over a repository of cultural objects through
a virtual maze where the objects are connected by pathways representing the
meaning relations over them. The approach of Labyrinth 3D leverages a system-
atic mapping of the conceptual model underlying the repository onto a virtual,
3D environment, to create an immersive and engaging experience for the user.
Designed to provide an alternative to the standard approaches to archive navi-
gation, Labyrinth 3D relies on the users’ curiosity to create personal paths in a
cultural domain.
In the next years, thanks to the advent of the paradigm of Linked Open
Data [44], semantically encoded information about cultural heritage, including
events, performances, collections, etc. will be available on the web from di↵erent
sources, enabling the experimentation of new paradigms in the presentation and
dissemination of cultural heritage. By applying the approach of Labyrinth 3D
to the design of new applications, it will possible to refine and improve the
approach described in this paper through practical case studies. The ultimate
goal of this research is to take full advantage of the whole range of the new
media languages, such as 3D, to develop creative and innovative applications in
the field of cultural heritage.
As a future work, we envisage the adoption of the software pipeline of
Labyrinth 3D in educational projects. In this setting, in fact, our assumption
is that the semantic-guided narrative exploration of themes, characters, epochs
etc. can provide a ludic path to knowledge access for young students and may
favour, in a gamification perspective, the process of knowledge acquisition. This
would require both an extension of the current catalogue of the archetypes as
22
well as an adaptation according to the specific needs of the educational project
considered.
8. Acknowledgements
The authors wish to thank Prof. Giulio Lughi for inspiration and discussion.
Our thanks go also to Neos s.r.l. for bringing to the Labyrinth project their
insights and contributions.
9. References
[1] S. Keene, Digital collections, Routledge, 2012.
[2] A. Padilla-Meléndez, A. R. del Águila-Obra, Web and social media usage
by museums: Online value creation, International Journal of Information
Management 33 (5) (2013) 892–898.
[3] N. Proctor, Digital: Museum as platform, curator as champion, in the age
of social media, Curator: The Museum Journal 53 (1) (2010) 35–43.
[4] W. van Hage, N. Stash, Y. Wang, L. Aroyo, Finding your way through the
rijksmuseum with an adaptive mobile museum guide, The Semantic Web:
Research and Applications (2010) 46–59.
[5] I. Horrocks, P. F. Patel-Schneider, F. Van Harmelen, From shiq and rdf
to owl: The making of a web ontology language, Web semantics: science,
services and agents on the World Wide Web 1 (1) (2003) 7–26.
[6] E. Hyvönen, Semantic portals for cultural heritage, in: Handbook on on-
tologies, Springer, 2009, pp. 757–778.
[7] M. Doerr, Ontologies for cultural heritage, in: Handbook on Ontologies,
Springer, 2009, pp. 463–486.
[8] B. Haslhofer, A. Isaac, data. europeana. eu: The europeana linked open
data pilot, in: International Conference on Dublin Core and Metadata
Applications, 2011, pp. 94–104.
[9] E. Hyvönen, J. Tuominen, E. Mäkelä, J. Dutruit, K. Apajalahti, E. Heino,
P. Leskinen, E. Ikkala, Second world war on the semantic web: The
warsampo project and semantic portal, in: Proceedings of 14th Interna-
tional Semantic Web Conference, 2015.
[10] M. Doerr, S. Gradmann, S. Hennicke, A. Isaac, C. Meghini, H. van de Som-
pel, The europeana data model (edm), in: World Library and Information
Congress: 76th IFLA general conference and assembly, 2010, pp. 10–15.
[11] M. M. Hall, From searching to using: Making sense of digital cultural
heritage collections, in: Proc. of “The Search is Over!” Exploring Cultural
Collections with Visualization, 2014.
23
[12] M. Nitsche, Video game spaces: image, play, and structure in 3D game
worlds, MIT Press, 2008.
[13] C. Ziemkiewicz, R. Kosara, The shaping of information by visual
metaphors, Visualization and Computer Graphics, IEEE Transactions on
14 (6) (2008) 1269–1276.
[14] A. Warburg, Der Bilderatlas Mnemosyne, Vol. 1, Akademie Verlag, 2008.
[15] S. Thompson, Myths and folktales, The Journal of American Folklore
68 (270) (1955) 482–488.
[16] E. Hyvönen, E. Mäkelä, T. Kauppinen, O. Alm, J. Kurki, T. Ruotsalo,
K. Seppälä, J. Takala, K. Puputti, H. Kuittinen, et al., Culturesampo: A
national publication system of cultural heritage on the semantic web 2.0,
The Semantic Web: Research and Applications (2009) 851–856.
[17] C. van den Akker, M. van Erp, L. Aroyo, R. Segers, L. Van der Meij,
S. Lgene, S. G., Understanding objects in online museum collections by
means of narratives, in: Proc. of the Third Workshop on Computational
Models of Narrative (CMN12).
[18] W. R. Van Hage, V. Malaisé, R. Segers, L. Hollink, G. Schreiber, Design
and use of the simple event model (sem), Web Semantics: Science, Services
and Agents on the World Wide Web 9 (2) (2011) 128–136.
[19] N. Gershon, W. Page, What storytelling can do for information visualiza-
tion, Communications of the ACM 44 (8) (2001) 31–37.
[20] J. Bruner, The narrative construction of reality, Critical inquiry (1991)
1–21.
[21] P. Mulholland, T. Collins, Using digital narratives to support the collabora-
tive learning and exploration of cultural heritage, in: Database and Expert
Systems Applications, 2002. Proceedings. 13th International Workshop on,
IEEE, 2002, pp. 527–531.
[22] P. Mulholland, T. Collins, Z. Zdrahal, Story fountain: intelligent support
for story research and exploration, in: Proceedings of the 9th international
conference on Intelligent user interfaces, ACM, 2004, pp. 62–69.
[23] P. Mulholland, A. Wol↵, T. Collins, Curate and storyspace: an ontology
and web-based environment for describing curatorial narratives, in: The
Semantic Web: Research and Applications, Springer, 2012, pp. 748–762.
[24] H. Kern, J. Saward, Through the Labyrinth: designs and meanings over
5000 years, Prestel New York, 2000.
[25] P. Rosenstiehl, Labirinto, Enciclopédia Einaudi 13 (1988) 247–273.
24
[26] K. Dylla, B. Frischer, P. Mueller, A. Ulmer, S. Haegler, Rome reborn 2.0:
A case study of virtual city reconstruction using procedural modeling tech-
niques, Computer Graphics World 16 (2008) 25.
[27] L. Calori, C. Camporesi, A. Negri, S. Pescarin, Virtual rome, in: SIG-
GRAPH Posters, ACM, 2008, p. 101.
[28] A. Guidazzoli, L. Calori, F. D. Ponti, T. Diamanti, S. Imboden, A. Mauri,
A. Negri, G. B. Cohen, S. Pescarin, M. Liguori, Apa the etruscan and 2700
years of 3d bologna history, in: SIGGRAPH Asia 2011 Posters, ACM, 2011,
p. 2.
[29] A. Lieto, R. Damiano, Building narrative connections among media ob-
jects in cultural heritage repositories, in: H. Koenitz, T. I. Sezen, G. Ferri,
M. Haahr, D. Sezen, G. Catak (Eds.), Interactive Storytelling - 6th Inter-
national Conference, ICIDS 2013, Istanbul, Turkey, November 6-9, 2013,
Proceedings, Vol. 8230 of Lecture Notes in Computer Science, Springer,
2013, pp. 257–260.
[30] R. Damiano, V. Lombardo, A. Lieto, Visual metaphors for semantic cul-
tural heritage, in: Intelligent Technologies for Interactive Entertainment
(INTETAIN), 2015 7th International Conference on, IEEE, 2015, pp. 100–
109.
[31] R. Damiano, A. Lieto, Ontological representations of narratives: a case
study on stories and actions, in: M. A. Finlayson, B. Fisseni, B. Löwe,
J. C. Meister (Eds.), 2013 Workshop on Computational Models of Nar-
rative, CMN 2013, August 4-6, 2013, Hamburg, Germany, Vol. 32 of OA-
SICS, Schloss Dagstuhl - Leibniz-Zentrum fuer Informatik, 2013, pp. 76–93.
doi:10.4230/OASIcs.CMN.2013.76.
URL http://dx.doi.org/10.4230/OASIcs.CMN.2013.76
[32] P. Lyman, B. Kahle, Archiving digital cultural artifacts, D-Lib Magazine
4 (7).
[33] L. Aroyo, N. Stash, Y. Wang, P. Gorgels, L. Rutledge, Chip demonstra-
tor: Semantics-driven recommendations and museum tour generation, The
Semantic Web (2007) 879–886.
[34] C. Van Den Akker, S. Legêne, M. Van Erp, L. Aroyo, R. Segers, L. van
Der Meij, J. Van Ossenbruggen, G. Schreiber, B. Wielinga, J. Oomen,
et al., Digital hermeneutics: Agora and the online understanding of cultural
heritage, in: Proceedings of the 3rd International Web Science Conference,
ACM, 2011, p. 10.
[35] G. Highet, The classical tradition: Greek and Roman influences on Western
literature, Oxford University Press, USA, 1949.
[36] E. T. O’Neill, Frbr: Functional requirements for bibliographic records, Li-
brary resources & technical services 46 (4) (2002) 150–159.
25
[37] D. C. M. Initiative, et al., Dublin core metadata element set, version 1.1.
[38] D. A. Norman, A↵ordance, conventions, and design, Interactions 6 (3)
(1999) 38–43.
[39] S. Burigat, L. Chittaro, Navigation in 3d virtual environments: E↵ects of
user experience and location-pointing navigation aids, International Jour-
nal of Human-Computer Studies 65 (11) (2007) 945–958.
[40] G. A. Miller, The magical number seven, plus or minus two: some limits
on our capacity for processing information., Psychological review 63 (2)
(1956) 81.
[41] R. Damiano, V. Lombardo, Labyrinth 3d. cultural archetypes for exploring
media archives, Digital Creativity.
[42] L. P. Rieber, D. Noah, Games, simulations, and visual metaphors in ed-
ucation: antagonism between enjoyment and learning, Educational Media
International 45 (2) (2008) 77–92.
[43] J. Oomen, L. Aroyo, Crowdsourcing in the cultural heritage domain: oppor-
tunities and challenges, in: Proceedings of the 5th International Conference
on Communities and Technologies, ACM, 2011, pp. 138–149.
[44] C. Bizer, T. Heath, T. Berners-Lee, Linked data-the story so far, Inter-
national journal on semantic web and information systems 5 (3) (2009)
1–22.
26