<!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>RITE: Rich Information Television</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Author Keywords Interactive TV</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Second Screen</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>User Experience</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Social Interaction</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Roman Ganhör Vienna University of Technology Favoritenstrasse 9-11</institution>
          ,
          <addr-line>1040 Vienna</addr-line>
        </aff>
      </contrib-group>
      <pub-date>
        <year>2015</year>
      </pub-date>
      <abstract>
        <p>When the Internet became more interactive it turned passive consumers of information into active providers of such. Over the time, volunteers built impressive databases filled with information for virtually every niche. In this paper we present RITE, an idea for augmenting films and TV serials with supplementary information that is freely available on the Internet. An easy, yet powerful interface and interaction design should encourages laid-back TV consumers to contribute. Thus, the system shall be optimized for different input devices such as tablets, smartphones or TV remotes. Besides, various automatism are intended to simplify the process of contribution even more.</p>
      </abstract>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>While Wikipedia is a universal dictionary other online
resources are more focused. The international movie
database (IMDb.com), for example, is a single purpose
database concentrating solely on motion pictures. The
website collects and provides information about movies as
well as TV serials. This information covers virtually every
aspect of a given motion picture including actors,
producers, dates, places, quotes and cross references to
other motion pictures.</p>
      <p>Whenever watching a motion picture on television or on the
Internet, the odds are good that a plethora of additional
information and metadata is available online. However, this
supplementary information is provided almost solely as text
and omitting opportunities for rich and appropriate user
experiences. Thus, retrieving additional information
requires a contextual change for the user, switching from a
passive viewer of a motion picture to an active investigator
for text. This contextual change not only disturbs the
viewing experience, the collected metadata is also
shortlived as there is no defined way or place to store it or share
it with other users.</p>
      <p>To overcome these limitations we propose RITE, a system
with an intuitive and efficient user interface and elaborated
interaction design allowing people to process, collect,
maintain and share metadata for any kind of film or TV
serials. The metadata is collected automatically or manually
from external online sources and put into relation to a given
movie or TV serial.</p>
      <p>BACKGROUND: TV AND INTERNET TECHNOLOGY
The habit of consuming television has widened from
scheduled programs on TV sets to video on demand
services displayed on TV sets, desktop computers and
mobile devices. Consequently, new features enriching the
TV experience have to consider the variations in means of
media transportation and consumption.</p>
      <p>
        Web techniques such as HTML and JavaScript can serve as
a common denominator. Internet based television embeds
HTML meta-information in the video stream or sends the
meta-information as a separate data stream, whereas
broadcast television (via satellite, cable, terrestrial) utilizes
HbbTV [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ] to embed HTML meta-information.
      </p>
      <p>Permitting users easy access to semi-structured data on
various websites like IMDb or Wikipedia the data has to be
pre-processed and structured using API's provided by the
websites and heuristics based on user behavior.</p>
      <p>Easing the annotation process even more advanced
techniques like face-recognition can be utilized. Plus, since
the numbers of persons acting in a film is limited and due to
external databases known to the algorithm the task of face
recognition can be enhanced in quality.</p>
      <p>RITE - RICH INFORMATION TELEVISION
In this section we describe the idea for a rich information
television system, RITE, in detail. We envision an
enhanced television experience based on additional in-situ
information. The information is provided by various
external or third-party Internet resources and can be
retrieved comfortably by a user while watching a movie or
TV serial. That is, whenever a user picks an actor on the
screen (with an appropriate device), RITE finds the name
for this actor and provides links to further information. Or
else, whenever a user queries information about a specific
building he or she has seen on the film RITE provides its
name and its location in-situ. Again, the information
provides links to further more detailed information.
As users who participate in retrieving and linking
information are crucial to the project, RITE has low entry
barriers and easy-to-use interfaces allowing even casual
users to contribute. The low entry barriers refer to the
organizational structure and concepts like editing and
contributing without prior registration. A transparent review
and feedback mechanism ranks contributions according
their quality and an undo-function reverts vandalism
entries. Another important aspect is the interface and
interaction design, addressing both the demands of novice
users and claims of advanced users. The challenges for a
feasible interface and interaction design are manifold,
however, in a first step we focus on three crucial tasks:
information aggregation and linking, information
maintenance and information invoking.</p>
      <p>
        Information Aggregation and Linking
An important task is finding relevant information and
linking this information to a film or TV serial. The process
of aggregating and linking additional metadata can be partly
automatized. So can sources like IMDb.com, Wikipedia and
other (not only film related) websites be crawled with
specialized programs and crawlers to retrieve relevant
information. As relevant information on various websites
can be structured or unstructured, the information can either
be gathered through specific API requests or must be
retrieved through parsing and semantic matching.
A starting point for automatized information gathering can
be online TV journals providing name and date of the
current TV program. Based on this initial information
further search queries can be conducted. Advanced
recognition systems such as reverse pictures search can
check results against the databases mentioned above and
enhance the overall quality of the automatized output.
However, in a second step a user manually confirms or
denies the automatically collected information. Depending
on the users judgment the information is ranked as
trustworthy or not trustworthy. The automatic process tries
to provide highly accurate information, however, the user
acts as a last instance. Thus, the interface design must be
both easy to understand and able to encourage even
laidback users to contribute [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ].
      </p>
      <p>The lowest level of user contribution is confirming or
denying information that was automatically retrieved and
provided by the system. In contrast, the highest level of
contribution is manually searching for information on the
Internet and linking this information to a specific scene in
the film or TV serial. To support users at this level the user
interface has to support video browsing and video
annotation. As actors or buildings are not visible throughout
the whole film, annotations start at some position in the
film and end at another position. Furthermore, the annotated
object's size can vary over time, due to zooming or
movement of the object. Automatized object tracking can
support users annotation tasks.</p>
      <p>Information Maintenance
Maintaining the information is also an important aspect to
keep metadata up to date and the system useful for the user.
RITE has to take care of various dimensions of
maintenance, such as keeping up-to-date with modifications
on external sources and overhauling existing information.
When changes to the external sources affects only the data
layer the changes can be incorporated automatically.
Whenever external sources have changes on their technical
base, such as the database layer, RITE must be adapted to
the new technical conditions.</p>
      <p>Overhauling existing information requires easy interfaces
and interaction design to animate lean back television
consumer to fix broken or false data.</p>
      <p>
        Information Invoking - Watching with RITE
In contrast to information aggregation, linking and
maintenance, information invoking is the unobtrusive and
optional task to invoke information during watching a
movie or TV serial. Depending on the device (tablet,
smartphone, TV remote) the interaction design is tailored to
their favorable affordances [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ].
      </p>
      <p>When using a tablet or smartphone the respective screen
replicates the content of the main screen and the touch
mechanism is utilized to provide a pleasant and easy user
experience. The interaction design for a TV remote
concentrates on the directional pad to allow slackened
usage even during dark conditions.</p>
      <p>USER INTERFACE
The user interface of RITE has to support two distinct
usecases. Whereas most of the time a user retrieves
information from the system, sometimes a user likes to add
information to the system.</p>
      <p>Figure 1 depicts a potential user interface for retrieving
information while a movie is running. The pink rectangles
indicate persons or things RITE has already information
stored about, like real name and film character name. The
purple rectangle, in contrast, mark persons or things other
user has marked as potentially interesting. In Figure 1 the
actor in the scene is selected and the information about him
is listed in the left section of the screen. With a remote
control a user can "hop" from one rectangle to another
utilizing the directional pad. When using a touch interface a
user touches within the boundaries of a rectangle to activate
the information bar.</p>
      <p>The amount of information within the information bar is
adapted to the screen size and can be adapted to the users
preferences.</p>
      <p>CONCLUSION
RITE should augment existing movies and TV serials with
additional meta-information that is freely available on the
Internet. To support user involvement, the interface and
interaction design will be adjusted to laid-back consumers.
In addition, advanced algorithms can be utilized to simplify
the users’ burden.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <given-names>Amazon</given-names>
            <surname>Corp</surname>
          </string-name>
          .,
          <year>2015</year>
          , http://www.alexa.com/topsites
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <surname>Institut für Rundfunktechnik GmbH - Open IPTV Forum</surname>
          </string-name>
          .
          <year>2015</year>
          . https://www.hbbtv.org/
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <surname>Knauf</surname>
            ,
            <given-names>R.</given-names>
          </string-name>
          et al.
          <year>2010</year>
          .
          <article-title>Constraints and simplification for a better mobile video annotation and content customization process</article-title>
          .
          <source>Workshop Proc. EuroITV.</source>
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <surname>Cruickshank</surname>
            ,
            <given-names>L.</given-names>
          </string-name>
          et.al.
          <year>2007</year>
          .
          <article-title>Making interactive TV easier to use: Interface design for a second screen approach</article-title>
          .
          <source>The Design Journal</source>
          <volume>10</volume>
          , no. 3
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>