<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>CultureCam: Exploring Europeana Images to Inspire Creative Design</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Sergiu Gordea Sanna Martilla</string-name>
          <email>1220, Vienna, Austria &lt;first&gt;.&lt;surname&gt;@aalto.fi &lt;first&gt;.&lt;surname&gt;@ait.ac.at</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Christina Holm</string-name>
          <email>&lt;first&gt;@spildaftid.dk</email>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="editor">
          <string-name>Creative Design, Image Similarity Search, Cultural Heritage</string-name>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>AIT-Austrian Institute of Aalto University, Technology GmbH Miestentie 3</institution>
          ,
          <addr-line>Donau-City-Strasse 1 Helsinki</addr-line>
          ,
          <country country="FI">Finland</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>Spild af Tid</institution>
          ,
          <addr-line>Copenhagen</addr-line>
          ,
          <country country="DK">Denmark</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>This paper presents an online tool that o ers designers, artists and all creative people, the possibility of exploring and accessing Europeana in a funny, playful and intuitive way. By using the webcam to take a photo of any object or texture, users can invoke the Culture Cam tool for accessing a set of images that are similar in color, shape or pattern. It is used to stimulate the inspiration of creative designers by diving into common cultural heritage and providing access to beautiful content available to be used as source for new derivative designs and art artefacts. The collection of items used by Culture Cam are accessible under public domain licence.</p>
      </abstract>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>1. INTRODUCTION</title>
      <p>The European digital library aggregates information from
cultural heritage institutions from the whole old continent.
Among those, there are more than 23 million images
representing for example traditional or modern artwork,
photographs of buildings and landscapes, scans of biodiversity
material or images from old manuscripts1.</p>
      <p>The aim of the CultureCam application is to connect the
digital cultural heritage made available by Europeana to
the open communities of craft and media designers. The
designers may use cultural heritage objects as sources for
new derivative designs such as embroidery, textile patterns,
3D printed objects, media art, etc. However, there are
important constraints with regard to the reusability of digital
content in new designs. The primary concerns are related
to the copyright protection and the quality of the content.
While Europeana has made e orts to assign usage rights to
all objects, only a small part of them are freely available
for reuse. From the reusable ones, some do not present
objects that raise the interest of creative designers. Typically,
they are only interested to use fascinating shapes, patterns
and/or color-rich content.</p>
      <p>The second challenge is related to the provision of e cient
and e ective ways for accessing inspiring images relevant for
1http://europeana.eu/portal/search.html?query=
TYPE:IMAGE&amp;rows=24
the task at hand. Often in their work, the designers are using
copper wires, paper forms or existing objects to sketch or
model new artifacts. They produce several versions of their
designs until the desired work is created. The Europeana
artifacts have a great potential to help and inspire designers
in their daily work; however, their concrete requirements
are hard to be described in words and the classic text-based
search engines are often not able to satisfy them. From
designer's point of view, accessing inspiring and reusable
content through Europeana portal2 is a complex and time
consuming task.</p>
      <p>The main motivation behind the development of
CultureCam is the need to address content accessibility issues
synthesized within the following questions:</p>
      <p>How to access and reuse Europeana content in an
interactive and playful way?
What kind of content can I nd in Europeana?
How can one nd object with particular visual
characteristics?
How to get motivated for exploring more image
content?</p>
      <p>How to access inspirational content for creative reuse?
In order to address these challenges, CultureCam employs
the search by example approach and interactive browsing of
a curated dataset. With the assistance of professional
designers we selected a set of reusable images from the featured
collections in Europeana Labs3, which is accessible through
CultureCam, both from desktop an mobile devices.</p>
      <p>The rest of the the paper describes the steps followed to
implement the CultureCam application and is organized as
follows: Section 2 places the current work within the related
research context of the cultural heritage domain. The
architecture and the process of building the CultureCam
prototype and the image processing and search functionality is
described in Section 3, while the graphical user interface is
presented in Section 4 followed by the discussion of
weaknesses and open issues in Section 5. The conclusion and
directions for future work are summarized in the nal part
of the paper.
2see http://www.europeana.eu/portal
3http://labs.europeana.eu/data/</p>
    </sec>
    <sec id="sec-2">
      <title>RELATED WORK</title>
      <p>
        The work presented in this paper is carried out within
the scope of Europeana Creative4 project. The project
follows a co-creation approach for developing ve pilot
applications empowered by the cultural heritage content accessible
through Europeana. A similar approach was employed for
the development of the CoDICE (COdesigning DIgital
Cultural Encounters) tool, as reported by Diaz et al in [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ]. This
has the goal to support co-designing activities for building
smart objects for enhanced encounters with cultural
heritage. The key functionalities of this tool are related to the
management and tracking co-design activities, enhancement
and documentation of design products.
      </p>
      <p>
        An important issue related to the reuse of cultural
heritage content in creative designs is caused by the copyright
restrictions. Many of the digitized cultural objects are not
freely reusable or they are available in low quality.
Martilla and Hyyppa show how copyrights become a driver of
the design process and override goals of human-centered and
participatory design [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ]. Rights clearance is extremely hard
in the case of audio-visual content. However, in the case
of image content, the identi cation of right owners is less
complicated, and the rights statements are available in the
metadata of Europeana objects. Still, less than ten percent
of image content is marked as public domain, and most of it
doesn't present much value for designers (i.e. large part of
public domain content are grayscale photographs of regular
landscapes, persons or objects).
      </p>
      <p>
        The core of the image similarity search engine used in this
paper was described by Amato et al in [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ]. A preliminary
evaluation of the service using a controlled dataset selected
from Europeana was presented in [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ]. The previous work
aimed at evaluating to which extent the search engine may
satisfy the expectations of the regular users or creative
design professionals. In contrast to the previous work, the
CultureCam dataset aggregates items with a large variation
of objects, patterns and shapes. There are only a few images
being very close to each example in terms of visual similarity.
      </p>
    </sec>
    <sec id="sec-3">
      <title>CULTURECAM PROTOTYPE</title>
      <p>The main goal of the Culture Cam prototype is to
provide easy and intuitive access to cultural heritage content
for designer communities. Therefore, the design goal of the
application is to provide an simple and intuitive mechanism
for accessing image content, based on a word free search and
a tactile interaction with the GUI. It invites the designers
to take a closer look into Europeana content by facilitating
easy access to the archive through provision of inspirational
and \unexpected" search results. Therefore, the search
module makes a tradeo between the similarity and serendipity
of the retrieved image content.
4http://pro.europeana.eu/web/europeana-creative
From the user perspective, CultureCam is a digital \live"
tool that facilitates the access to images containing similar
colors, shapes and patterns by using a web camera. One may
take a snapshot of an object in front of his/hers computer or
mobile device and submit it to Culture Cam, which analyzes
the given picture and comes up with a result set of similar
content.</p>
      <p>
        Figure 1 presents the architecture of the proposed system,
which allows end users to use their favorite mobile devices
(phones or tablets) or computers (notebooks or desktops)
for accessing the frontend (web) application. The interactive
access and browsing of images is empowered by the search
functionality provided by the backend service in charge of
managing the image index.
The process of selecting and indexing appropriate
materials for being reused by designers is not a straight-forward
one. The approach used for accomplishing this goal is
presented in Figure 2, following a three-step work ow. In the
rst step, a manual pre-selection of available collections and
items is performed. This makes use of expert knowledge
to search content basing on existing metadata followed by
a brief visual inspection of the given selection. For the
CultureCam prototype we primarily investigated collections
containing public domain content, which are available also
with good resolution and a wide variation of colors, patterns
and shapes. The visual inspection is made directly within
the Europeana Portal 5. In the second step, the targeted
images are accessed through the Europeana API 6, the image
features are extracted and images that doesn't pass a certain
level of quality are ltered out (including placeholders, small
resolution, grayscale or monochromatic images). Finally, the
images are evaluated by their visual features and placed in
an image index. A nearest neighborhood algorithm is used
here for reducing the search space and improve the
execution performance at runtime. This is achieved by selecting
a so called pivot set and computing the distances between
each image and each pivot. At runtime, the search results
are ordered by their similarity relative to the pivot set. A
detailed description of the feature extraction and indexing
process is available in [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ].
4.
      </p>
    </sec>
    <sec id="sec-4">
      <title>GRAPHICAL USER INTERFACE</title>
      <p>Due to the complexity of accessing and browsing the
Europeana repository, a clear goal of Culture Cam is to make the
interaction with the graphical user interface as simple
intuitive and engaging as possible. Consequently, the graphical
user interface is remarkably simple, the number of
interaction possibilities and features in the tool being reduced to
a minimum. The understated design aims to create a
subtle frame for accessing beautiful Europeana content,
highlighting images and thus presenting the search result as nice
5see http://www.europeana.eu/portal/
6see http://labs.europeana.eu/api/
gallery pages. Additionally, it has been the intention all the
way through to keep the user journey short. There are no
conventional menu/sub menus structures except from the
menu icon, borrowed from the iPhone interface with the
clear purpose of creating a simple, intuitive and \familiar"
user interface.
The rst screen of the application invites the user to
capture a photo that will automatically trigger a search in the
image repository as shown in Figure 3. Therefore, the GUI
uses no more than three control elements for taking the
picture and for back and forth navigation through the search
results. Recently, new functionality for sharing the links to
the current search results on Facebook and Twitter has been
added to this screen. The latest version of the CultureCam
can be accessed on the o cial website7.
When users choose to take a close look at an image
retrieved from the Europeana repository, four more controls
are displayed. One of them is used to display basic metadata
of the image (i.e. title, rights, provider) and another
represents the direct link to Europeana page providing complete
information about the given image. The other two
possibilities are to go back and take a new picture or dig further
into Europeana content by searching similars to the current
selection.</p>
    </sec>
    <sec id="sec-5">
      <title>OPEN ISSUES</title>
      <p>The success of the CultureCam application is completely
dependent on the quality and the amount of content made
exposed thought its image search service. Selecting the
appropriate items for being included in the CultureCam
collection is a complex process, as described in Section 3. Manual
curation is not an option when the size of the collection
increases to to thousands of objects or more. It is also not an
7http://culturecam.eu
scalable or repeatable/reproducible process. Therefore, we
employed several heuristics for semi-automatic content
selection by using the resolution, color distributions and shape
descriptions in order to identify the most interesting content
items. Another important policy in the content selection
process was introduced by the need of including only
content that is freely reusable for design purposes. Currently
we only included public domain8 content, in the next
versions we plan to include also images released under the CC09
license.</p>
      <p>Capturing photos with the webcams integrated in
computers or mobile devices is the core functionality of the
CultureCam prototype. Webcams are available in the great
majority of notebooks and mobile phones, however, their
heterogeneity is transposed in the technical and visual
characteristics of the captured photos. In the most of the cases,
the images are not captured under optimal illumination
conditions and as a consequence they are undersaturated, grainy
or dark. In contrast, the Europeana images are created by
scanning with professional devices. Often, the objects are
placed on background surfaces of di erent colors. An
additional step for processing the input images might be
implemented in the future to overcame the systematic di erences
between the captured images and Europeana content and
thus, further improve the search results.
6.</p>
    </sec>
    <sec id="sec-6">
      <title>CONCLUSIONS AND FUTURE WORK</title>
      <p>This paper presents a tool that makes the visual
navigation through the selected image content available in
Europeana as simple and intuitive as possible. The current
version of the CultureCam uses a collection of 2600 items,
that were curated to be freely reusable and also inspiring for
creative designers. When being focused on simplifying the
browsing and stimulating the interest for cultural heritage
items, the search algorithm was designed to make a
tradeo between the similarity of the retrieved results and their
serendipity. At the submission date, the CultureCam tool
is in an pre-release version. Future work will investigate the
possibility to provide access to whole image content of
Europeana and to evaluate the acceptance of the CultureCam
application within the designer communities.
7.</p>
    </sec>
    <sec id="sec-7">
      <title>ACKNOWLEDGEMENTS</title>
      <p>This work was partially funded by the EuropeanaCreative
project, co-funded by the Commission of the European
Communities under the ICT Policy Support Programme (ICT
PSP). This publication re ects only the authors' views and
the European Union is not liable for any use that might be
made of information contained therein.
8.
8http://creativecommons.org/publicdomain/mark/1.0/
9http://creativecommons.org/publicdomain/zero/1.0/</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          [1]
          <string-name>
            <given-names>G.</given-names>
            <surname>Amato</surname>
          </string-name>
          ,
          <string-name>
            <given-names>P.</given-names>
            <surname>Bolettieri</surname>
          </string-name>
          ,
          <string-name>
            <given-names>F.</given-names>
            <surname>Falchi</surname>
          </string-name>
          ,
          <string-name>
            <given-names>C.</given-names>
            <surname>Gennaro</surname>
          </string-name>
          , and
          <string-name>
            <given-names>F.</given-names>
            <surname>Rabitti</surname>
          </string-name>
          .
          <article-title>Combining local and global visual feature similarity using a text search engine</article-title>
          .
          <source>In CBMI</source>
          , pages
          <volume>49</volume>
          {
          <fpage>54</fpage>
          ,
          <year>2011</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          [2]
          <string-name>
            <surname>P.</surname>
          </string-name>
          <article-title>D az, I. Aedo, and</article-title>
          <string-name>
            <given-names>J.</given-names>
            <surname>Cubas</surname>
          </string-name>
          . Codice:
          <article-title>Balancing software engineering and creativity in the co-design of digital encounters with cultural heritage</article-title>
          .
          <source>In Proceedings of the 2014 International Working Conference on Advanced Visual Interfaces</source>
          ,
          <source>AVI '14</source>
          , pages
          <fpage>253</fpage>
          {
          <fpage>256</fpage>
          , New York, NY, USA,
          <year>2014</year>
          . ACM.
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          [3]
          <string-name>
            <given-names>S.</given-names>
            <surname>Gordea</surname>
          </string-name>
          .
          <article-title>An image similarity search for the european digital library and beyond</article-title>
          . In L.
          <string-name>
            <surname>Bolikowski</surname>
            ,
            <given-names>V.</given-names>
          </string-name>
          <string-name>
            <surname>Casarosa</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          <string-name>
            <surname>Goodale</surname>
            ,
            <given-names>N.</given-names>
          </string-name>
          <string-name>
            <surname>Houssos</surname>
            ,
            <given-names>P.</given-names>
          </string-name>
          <string-name>
            <surname>Manghi</surname>
          </string-name>
          , and J. Schirrwagen, editors,
          <source>Theory and Practice of Digital Libraries - TPDL 2013 Selected Workshops - LCPD</source>
          <year>2013</year>
          ,
          <article-title>SUEDL 2013</article-title>
          ,
          <article-title>DataCur 2013</article-title>
          , Held in Valletta, Malta,
          <source>September 22-26</source>
          ,
          <year>2013</year>
          .
          <source>Revised Selected Papers</source>
          , volume
          <volume>416</volume>
          of Communications in Computer and Information Science, pages
          <volume>190</volume>
          {
          <fpage>196</fpage>
          . Springer,
          <year>2013</year>
          .
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          [4]
          <string-name>
            <given-names>S.</given-names>
            <surname>Marttila</surname>
          </string-name>
          and
          <string-name>
            <given-names>K.</given-names>
            <surname>Hyyppa</surname>
          </string-name>
          <article-title></article-title>
          . Rights to remember?:
          <article-title>How copyrights complicate media design</article-title>
          .
          <source>In Proceedings of the 8th Nordic Conference on Human-Computer Interaction: Fun</source>
          , Fast, Foundational,
          <source>NordiCHI '14</source>
          , pages
          <fpage>481</fpage>
          {
          <fpage>490</fpage>
          , New York, NY, USA,
          <year>2014</year>
          . ACM.
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>