|
|
VITALAS User Manual (v1.5 – Feb. 2010) |
|
|
VITALAS User Manual (v1.5 – Feb. 2010) |
VITALAS is a use-case driven project that aims to deliver a pre-industrial prototype allowing intelligent access to large scale multimedia professional archives, thorough personalized services coupled with technological functionalities.
VITALAS is the Next Generation of Multimedia Search Engine that enables content-based retrieval and browsing of very large audio and visual datasets without pre-existing metadata, through different advanced technologies:
· Automatic retrieval of spoken words through advanced syllable-based speech recognition
· Automatic annotation through generic “object” recognition allowing post-editing of multimedia content
· Content-based filtering, re-ranking and browsing of visual contents.
· Automatic structuring of video streams.
· Automatic duplicates and near-duplicates retrieval
The VITALAS system technology is envisaged to be a B2B tool which develops innovative services, with a clear potential to be adopted by a larger public and to be reused by consumers demanding robust multimedia search engines. The main advantages of VITALAS over existing technologies are:
· its provision of automatic annotation functions
· its use of innovative interfaces for searching in large audio visual archives, and
· its capacity to perform personalised searches for each and every user
VITALAS offers personalization that allows the system to adapt to a user’s context of operation. Knowledge about the users and their goals is a key to retrieve the most relevant documents for a particular context. This results in a system that can be customised according to the needs of a particular user.
The authentication process is available at the VITALAS Welcome page (see next Fig.). The user must provide a user name and a password. The user can change his/her password by accessing the user’s profile at the VITALAS top menu.
Currently the users must be logged in for performing searches in VITALAS. The authentication mechanism will allow the user to exploit all system functionalities, since some of them are based on user’s identification, e.g. light boxes or personalization.
Users can register in an easy way by clicking on the “Register” button. In the registration form the user is requested to provide some data (see next Fig.). Some of the fields are mandatory (marked by ‘*’) and the user has to accept the ‘Terms of use’ clauses. When entering the company name, if this name already exists, the rest of the fields related to the company are automatically loaded on the page.
In the registration process a new user account is created and the user is assigned a default Role and a default Access Profile. Higher level of privileges can be assigned upon request to the system Administrator.
The information provided by the user at registration time can be further completed/modified by him/her by accessing the user’s profile at the VITALAS top menu.
Fig. 2 Registration Form
VITALAS system provides a common interface for searching image and video datasets. Some specific features have been developed for each media type, but the general search layout is identical.
The next figure shows a global view of the VITALAS Search page and Results presentation layout, which is composed of several sections that are explained in detail below.
Fig. 3 VITALAS general overview
Menu bar
At the top of the window the user can find a menu bar with the following components:
· VITALAS logo (link to the VITALAS home page)
· Name of the logged user
· Links to the main application modules (search, administration and my profile)
· Help information (video demo, help)
· Link to a questionnaire to get user feedback on the VITALAS project and tool
· Login/Log-out button
Search modes
Below the menu bar the user can find two tabs that provide a quick access to the “Basic Search” mode and the “Advanced Search” mode respectively.
In the figure the “Basic Search” tab is shown and within it the default layout of this type of search is displayed, including a query text box and two buttons: “Search” and “Reset”. The two icons at the right of the text box, allow the selection of the media type on which the search will be performed: images or videos.
Search Results
The result set area contains several tabs that allow a quick access to the main features of the results presentation module:
The result set area also contains some information related to the results, such as the total number of retrieved results, the navigation mechanism to go forward and backward among the result pages, the sorting criteria and the number of items per page.
We have prepared a series of video demos that explain main features of the system. They are available through the ‘video demo’ link on the header of the application. Have a look at them.
Generalities
Starting
with VITALAS
3. How to login? How to logout?
Searching
Images
10. Options for the search results
13. Interactive Cartographic Explorer
Searching
Videos
14. How to search videos? How to work with videos?
Being aware that professional users have different needs when formulating a query, the VITALAS system provides an adaptive interface for searching multimedia content:
· Basic search (query text box to input some keywords)
· Basic expanded search (basic search plus some extended options shown on user’s demand)
· Advanced search (advanced functionalities for complex queries)
Basic Search comprises just a query text box and two buttons: “Search” and “Reset” (see next Fig.). When the user types some keywords in the text box, an auto-completion facility helps him suggesting previous keywords. Logical operators (AND, OR, NOT) and brackets [ ] are allowed in the query box.
The two icons at the right of the text box, allow the selection of the media type on which the search will be performed: images or videos. The default value can be defined by the user in the Profile section.
The “x”
icon (
),
that is located close to query text box, allows to perform a partial reset on that field.
The Basic expanded search is displayed on user request by clicking on the button located at the top-right of the Basic Search page. As a result more search options are displayed allowing the user to refine the search (see next Fig.).
The “x”
icon (),
that is located close to the search options, always performs, when clicked, a partial reset on that option.
Fig. 5 Basic Expanded Search page
The Advanced Search option opens a dialog box where several metadata fields are displayed (see next figure). The user can then define a more precise query using these fields.
The user can select some metadata fields to compose the query. These metadata fields are linked by default by the AND operator. Any combination of logical operators (AND, OR, NOT) and brackets [ ] can be used.
To perform a search on the VITALAS Image datasets the user must click on the “images” icon, that is located at the left side of the “Search” button.
Fig. 6 Search in image datasets
When searching for images VITALAS supports different types of queries, which can be executed independently (pure search) or, most of them, can be combined to create a Fused query:
· Local Search (*)
· Relevance Feedback search (*)
(*) not available
for Fused search
In this type of search the user enters some keywords in the query text box and clicks on the “Search” button. The default operator in Full text search is ‘OR’ (Lucene default operator).
In the Basic Expanded Search window, before executing a query, the most popular keywords are displayed in the “Keywords” box. After some results are returned for the query, the system displays some relevant Keyword suggestions in that box.
For each item in the tag cloud, two ways of adding it to the query are provided: AND (when clicking in the ‘+’ sign before the item) and OR (when clicking on the item itself). Then, the user can launch a new query by clicking on the “Search” button.
Query example: car AND competition
In Concept search the user selects some multimedia concepts from a concept list and clicks on the “Search” button.
Before launching the query the most popular concepts are displayed in the “Concepts” box. After some results are returned for the query, the system displays some relevant Concepts suggestions on that box.
For each item in the tag cloud, two ways of adding it to the query are provided: AND (when clicking in the ‘+’ sign before the item) and OR (when clicking on the item itself). Then, the user can launch a new query by clicking on the “Search” button.
Query example: car AND [concept:competition]
In Metadata search (Advanced search) the user provides some specific values for certain metadata fields that will be added to the query search box (see next Fig.). These fields are grouped in two subsets: one for the most common used fields (Title, Description, Keywords, Creator, Creation Date, Creator Title) and the other subset for the rest of fields (Category, City, Language, etc). The query can be also edited by typing in the summary field (large blue background field at the top of the form).
Once the fields have been filled in, the user must click the “Add to search” button. The values provided by the user will be added to the query search box, then the user must click the ‘Search’ button.
Query example: car AND [title:Audi] AND [keywords:speed]
Fig. 7 Metadata search for images
In Similar search the user selects an image from a result set to launch a new search. The result of the search will be a set of images similar to the selected one (global visual similarity). Similar search can be launched from two points of the user interface:
· Click on a thumbnail (Zoom window); click the “~”button (see next Fig.)
· Mouse over a thumbnail (Mini-toolbar); click the “i” button (Details window); click the “~ similar” button
In Fused search, the similar search facility is supported by dropping-and-dragging a thumbnail into the similar box. After defining the rest of components for the fused query, the user must click the ‘Search’ button.
Fig. 8 Launch Similar search for images on the Zoom Window
In Colour search the user selects a colour from a colour picker control one in order to find results with this predominant colour. It is available in the Basic Expanded search.
Most basic colours are shown as direct selections, and a more detailed colour selection dialog can be opened if needed (see next Fig).
In this type of search the user provides some external image to search for similar images (global visual similarity). It is available in the Basic Expanded search.
When the user clicks on the “Upload” button she is requested to select an image from a local file (see next Fig.). The selected image is displayed in the box. Then, she can perform a similar search by clicking on the “~” button. Optionally she can also display some details of the uploaded image by clicking on the “i” button”.
Fig. 10 Search by image upload (two steps)
Local search is a special type of ‘Similar search’ that can used to search for images that are similar to a certain area of a selected image. Local search can be launched from the “Details window” (see next Fig.).
So, to perform a local search from a certain image the user will open the “Details window” (button “i” located both at the mini-toolbar and at the ‘Zoom window’). Then he will use the mouse to select the area of interest in the picture (purple rectangle) and click the “local similar” button.
Fig. 11 Local search in the Details window
The relevance feedback search is a special modality of ‘Similar search’ that can used to search for images that are similar to one or more relevant images selected by a user and/or have nothing in common with one or more non-relevant images also selected by the user.
Next figure shows how to start a Relevance feedback search from the Mosaic presentation view. The user has several options to mark an image as being Relevant (+ feedback) or Irrelevant (- feedback):
· Mouse over a thumbnail for some seconds (Mini-toolbar) and click on the ‘+’|’-‘ button
· Click on a thumbnail (Zoom Window), and click on the ‘+’|’-‘ button
· Open the “Details window” an click on the ‘+’ |’-‘ button
Relevant images will be surrounded by a green frame and irrelevant images with a red frame.
To launch a relevance feedback search the
user will click the 
button, located on the “my feedback” tab.
Several iterations are allowed; in each of them the user will select the
corresponding relevant/irrelevant images and will click on the button. The
numbers displayed on the tab correspond to the number of positive and negative
images at that iteration.
To start a new relevance feedback process,
i.e. remove all previous selections, the user will click the 
button, also located on the “my feedback” tab.
Fig. 12 Relevance feedback search
To check which items have been marked as relevant or irrelevant in the feedback iterations the user will click the “my Feedback” tab. As a result two baskets (+ and -) are displayed with its respective document thumbnails (see next Fig.).
The user can empty the positive basket, the
negative basket or both of them. Also, she can remove individual items from the
baskets by clicking on the 
button.
A toolbar is displayed at the bottom of each thumbnail containing the following actions:
· “+” and “-“: mark the document as positive or negative regarding relevance feedback
· “i”: displays the “Details window” for the document
·
“
”:
adds the document to the active light box
Fig. 13 Relevance feedback baskets
The Fused search consists on the combination of different types of queries to perform a search. The user has to make sure that the “Images” icon is selected.
To launch a Fused search the user must perform more than one of the following actions:
· Enter some keywords in the query text box
· Select one or more keywords from the keyword suggestion box
· Select one or more concepts form the concept suggestion box
· Open the Metadata form and provide values for some metadata fields
· Drag and drop a thumbnail from the result set to the Similar search box
· Select a colour in the colour picker
Then, the user must click on the “Search” button to start the search.
To perform a search on the VITALAS video datasets, first of all, the user has to click on the “Video” icon, which is located at the left side of the “Search” button.
Fig. 14 Search in video datasets
When searching for videos VITALAS supports different types of queries, which can be executed independently (pure search) or can be combined to create a fused query:
Audio Search allows the user to search for certain keywords within the audio track of the VITALAS video datasets. The user can provide several keywords and can link them by using a relational operator (AND, OR). Square brackets [ ] are allowed to group the query terms. After clicking the ‘search’ button and the relevant results are displayed in the results area.
In VITALAS, Audio search and Full text search are both supported by the same query box, being Audio search the default search type for video. The user could change this option in his profile (‘preferences for search’ section).
In Full text search the user enters some keywords in the query text box and/or selects some keyword suggestions and then clicks the “Search” button. The user can provide several keywords and can link them by using a relational operator (AND, OR). Square brackets [ ] are allowed to group the query terms.
In VITALAS, Audio search and Full text search are both supported by the same query box, being Audio search the default search type for video. So if the user wants to launch a full text search he will have to select this option in his profile (‘preferences for search’ section).
In Concept search the user selects some multimedia concepts from a concept list and clicks on the “Search” button. It is available in the Basic Expanded search.
Before launching the query the most popular concepts are displayed in the “Concepts” box. After some results are returned for the query, the system displays some relevant Concepts suggestions on that box.
For each item in the tag cloud, two ways of adding it to the query are provided: AND (when clicking in the ‘+’ sign before the item) and OR (when clicking on the item itself). Then, the user can launch a new query by clicking on the “Search” button.
In Metadata search (Advanced search) the user provides some specific values for certain metadata fields that will be added to the query search box (see next Fig.) and the system performs a search into the different metadata fields of the VITALAS video datasets.
The resulting query is displayed at the summary field (large blue background field at the top of the form). This field can be edited by the user to modify the query.
When the user clicks on the “Add to search” button the resulting query is added to the main search box. Then the user has to click the “Search” button.
In Similar search the user selects a keyframe from a result set to launch a new search. The result of the search will be a set of similar video segments. Similar search can be launched from two points of the user interface:
· Click on a thumbnail (Zoom window); click the “~”button (see next Fig.)
· Mouse over a thumbnail (Mini-toolbar); click the “i” button (Details window); select a keyframe from the video timeline and click the “~ similar” button (see section 5.2 Video details).
Fig. 16 Launch Similar search for videos on the Zoom Window
In Fused search, the similar search facility is supported by dropping-and-dragging a thumbnail into the “Similar search” box. After defining the rest of components for the fused query, the user must click the ‘Search’ button.
The Fused search consists on the combination of the previous types of queries to perform a composed query. The user has to make sure that the “Videos” icon is selected.
To perform a Fused search the user will perform more than one of the following actions:
· Enter some keywords in the query text box
· Select one or more keywords from the keyword suggestion box
· Select one or more concepts form the concept suggestion box
· Provide values for some metadata fields in the query text box
· Drag and drop a thumbnail from the result set to the Similar search box
Then, the user clicks on the “Search” button to start the search.
The VITALAS system provides common interfaces for displaying the results retrieved either from an image query search or from a video query search. When displaying video results each thumbnail corresponds to a matching segment.
Two presentation modes are available in VITALAS:
The details of each result are displayed in detail in two windows:
Finally, three advanced visualisation tools have been integrated:
· A Global Exploration Map (GEM)
· An Interactive Local Explorer (ILE)
· An Interactive Map of TV news content (NIM)
In the Mosaic View presentation mode, each query result is represented by a thumbnail. The results are displayed ordered by relevance and the number of items to be displayed in each page is automatically computed (see next Fig.).
The GUI provides three tools to interact with the results’ thumbnails: Mini-toolbar, Zoom Window and Hot keys.
The Mini-toolbar is displayed when the user holds the mouse on a thumbnail for a few seconds. As a result, a blue frame surrounds the thumbnail and a mini toolbar is displayed at the bottom of it (see next Fig.).
The buttons of the toolbar support the main operations to be performed on a result:
“+” and “-“: mark the result as positive or negative regarding relevance feedback
“i”: displays detailed information on the result
“”:
adds the result to the active light box
Fig. 18 Mini-toolbar (mouse over a thumbnail)
The Zoom Window is displayed when a user
clicks on a thumbnail, showing the main details of
the result (title and keywords).
If the result media type is Image, then the Zoom window shows a bigger size image. If the result media type is Video, then the Zoom window shows a sequence of the main key frames of the video. In both cases, some buttons are available to launch main operations:
“+” and “-“: marks the result as positive or negative regarding relevance feedback
“i”: displays detailed information on the result (‘Details window’)
“*”: for images, launches the Interactive Local Explorer. For videos, launches the Interactive Map of TV news content.
“~”: performs a similar search
“”:
adds the result to the active light box
The Zoom window can be closed by clicking the ‘x’ button. When the user clicks on another thumbnail the Zoom window changes showing the main details of the selected result.
Fig. 19 Zoom Window (click on a thumbnail)
The implementation of hot keys functionality allows professionals to perform repetitive tasks in a quick way (such as marking results to be positive or negative for the relevance feedback mechanism).
The box providing information on the 'Hot keys' is displayed when the user clicks on the "Hot keys" link, located at the left middle part of the screen in 'Mosaic view', 'Lightboxes' and 'MyFeedback' tabs. Cluster view does not support the Hot keys functionality.
Next figure shows the list of hot keys that are currently available in VITALAS:
For example, when the mouse is over a thumbnail and the user presses the key “p” then the result is marked in green.
Another example, when the mouse is over a thumbnail and the user presses the key “d” then the “Details Window” of the result is displayed.
The Cluster View gives the user more control over the search tool and, at the same time, displays the results of a complex query showing how the different sub-searches perform the query. The aim is to show if the good results are matching the text criteria/metadata criteria/concepts criteria/etc, and hence let the user fine-tune the engine to produce different results. For example, he can give more weight to the concepts matching criteria.
The results in Cluster View are not shown alltogether like in Mosaic View, but grouped by matching criteria:
The user can explore deeper into each individual group (pressing the “<” and “>” buttons on the top-right side), or close groups that are not informative (and this way revise the query for the next iteration).
The slider controls (left side of bar) on each specific group represent a way to control how much weight the user want to give to each one of the available criteria. This way, if the weight he gives to the Concepts criteria is increased, the results obtained in the Fused group (the first group, that matches multiple criteria) should change, showing results closer to the concepts used in the query.
The next figure shows the appearance of the result set presentation in the Cluster view.
In the cluster view, each result is also represented by a
thumbnail. The handling of the thumbnails is similar to the Mosaic View, except
that the Zoom View is displayed on mouse-over.
Light boxes allow the user to store specific documents resulting from a search process. They can hold images and videos, and are related to a specific user. Therefore, light boxes are available only when the user has logged into VITALAS.
When a user is registered into the VITALAS system, a default light box named “My Lightbox” is created as part of the profile data. Later on, the user will be able to change the name of this light box and will also be able to create new light boxes. Light boxes are stored at the server, associated to the user account and are reloaded in each new session.
A user can have multiple light boxes, but at each moment only one of them will be the active light box, this means that when the user selects an item in the search results and clicks the “to light box” button, the selected item will be added to the active light box.
The user can change the active light box using the ‘▼’ icon in the ‘Light box’ tab (see next Fig.). When clicking on it, a list of his light boxes is displayed, and the user can select one of them (a tooltip showing the description of each light box is displayed when the user moves the mouse over its name). The name of the newly selected light box is then displayed in the light boxes tab, as well as the number of items it contains.
Fig. 22 Change the active light box
To open the active light box in order to see and manage its contents, the user has to click on the light box name displayed at the tab (this opens the light boxes tab).
If the user wants to open a light box that is not the active light box, he has two different two-steps options:
A) First, set it as the active light box (clicking on the light box tab ‘▼’ icon and selecting its name) and, afterwards, click on his name on the light boxes tab. Or vice versa:
B) First, open the active light box (clicking on the light box name ) and, afterwards, select the required light box (clicking on the light box tab ‘▼’ icon and selecting it from the menu).
When a light box is opened, the documents stored
in it are displayed (see next Fig.). Each thumbnail has a delete button (
icon) that allows removing the document from the light box. A toolbar at the
bottom of each thumbnail includes the main actions to be performed with the
document:
“+” and “-“: mark the document as positive or negative regarding relevance feedback
“i”: displays detailed information on the document
Fig. 24 Contents of Light box ‘sandbox’
At the right of the window a list of actions is displayed showing the main operations that can be performed on light boxes. These are:
Manage “LightboxName”
Includes some actions to manage the items stored in the opened light box:
Manage Lightboxes
Includes actions to manage light boxes:
To close the active light box and get back to the results of a query, the user has to click on the ‘Mosaic View’ button.
The “Details Window” of an image result is displayed when the user
clicks on the “i” button, both at
the mini-toolbar and at the “Zoom Window”. This window shows a medium size
image and a set of metadata fields (see next Fig.). The user can define some
default metadata fields for the detailed view in his Profile. The “more
details” link will display all the metadata fields.
Fig. 25 Details window for an image
At the bottom of the image there is a toolbar that supports the main operations on the result:
“+” and “-“: mark the result as positive or negative regarding relevance feedback
“* ICE”: launches the Interactive Local Explorer
“local similar”: performs a local similar search, based on the area selected by the user
“~ similar”: performs a similar search
“
to lightbox”: adds the result to the
active light box
To further refine the search, the user can visually specify which concepts of the result are added to a new query (concepts are clicked in the ‘Concepts’ box and dragged in the ‘Add to search’ box). Then, the “search” button will be clicked to start a refined query.
The “Details Window” of a video result is displayed when the user
clicks on the “i” button, both at
the mini-toolbar and at the “Zoom Window”. This window shows a video player, the
video metadata, three timelines and the video concepts (see next Fig.).
Fig. 26 Details window for a video
The video player allows handling the video in a standard way: play
it back from the beginning or from any point selected by the user in any
timeline; pause it; stop it, etc. A QuickTime
player has been selected which allows playing RTSP stream videos. In order to
use the player, the appropriate plug-in needs to be installed in the browser.
At the bottom of the video player, there is a toolbar that supports
the main operations on the result:
“
to lightbox”: adds the result to the
active light box
Three sets of metadata fields are displayed at the right of
the video player in separate tabs -document metadata, video metadata and audio
metadata-.
The three timelines, displayed at the bottom of the video player, are
divided in segments to show the video structure and mark those segments
containing a matching result. The colour intensity of
each segment represents its score regarding the query.
Below the video timeline there is
a zoom utility that allows to see in more detail part of the whole timeline.
Buttons labelled as “1x”, “2x”,… “100x” provide different zoom views of the
video segments. The zoom can be scrolled by clicking on the arrows.
The timelines and the video player are synchronised, so that when
a user clicks on a certain segment, the video player reproduces the
corresponding video sequence. At the right of each timeline two arrow icons (left,
right) also provide facilities for moving through the segments and starting
video reproduction.
When a segment is clicked, its score and its metadata fields are
displayed in the ‘Segment metadata’ box, whereas the concepts related to the segment
are displayed in the ‘Segment concepts’ box.
To further refine the search, the user can visually select some concepts to be added to a new query (concepts are clicked in the ‘Concepts’ box and dragged in the ‘Add to search’ box). Then, the “search” button will be clicked to start a refined query.
This application is launched from the VITALAS interface by clicking the button “Advanced visualization” on the top of the result page.
The Global Exploration Map (GEM) presents the data set based on textual similarities. Documents that are semantically close from each other (sharing descriptive keywords) are positioned close to each other on the map creating semantically dense neighbourhoods (coloured clusters). This map can be zoom-in and zoom-out in the central menu to access either to the clusters level or to the images level.
Classical navigation functions are available:
· Rolling wheel to zoom in and out
· Mouse displacement maintaining the click button allows translating the map.
Two additional menus at the right and at the left of the graphic map are provided to help the user to navigate in the map:
The right panel gathers functional aspects:
· Search function: a classic textual input field is available on the top of right panel to enter queries, the function is launched when the user clicks on the search button. The results of the search process are highlighted on the map, shaded tones (red to yellow) are used to mark the rank of the results. Two modalities are available for the search: text or visual. Textual queries are entered directly by the user. To make a visual query the user should select a thumbnail on the map (with ctrl+ mouse click): the URI of the image is displayed on the search text field. Both query types may be combined.
· Parameter setting for results display: The number of results displayed may be given by the user in a text input field. (example of the action of this parameter on the map is shown above)
· Clusters topic overview: A list of the most descriptive terms corresponding to the cluster pointed by the mouse (+ ctrl function) allows the user to understand easily the clusters topics.
The left panel is used to help the user in the navigation:
· Overview: The zone displayed on the central panel is shown with a grey transparency square on reduced map of the left top panel (cf. fig 31 36)
· Document image view: displays the picture of the document currently pointed by the mouse (+ shift)
· Document text view: displays the textual metadata of the document currently pointed by the mouse (+ shift)
Fig. 27 Presentation of the cluster map
This tool allows the user to build local maps based on visual textual or mixed similarities. To launch it, the user has to perform an search and click on an image in the results. Then the ILE tool may be invoked by clicking on the ‘*’ button in the ‘Zoom’ Window or on the ‘* ICE’ button in the “Details” Window.
The image appears in the main panel and the user may explore the proximity by simply double clicking on the image. The ten nearest documents appears on the map, and are linked to the original image (see next Fig.). The user may choose the modality used to express the nearest neighbours: textual, visual and mixed (textual + visual) are available.
Navigation functions (zoom and translation) of the Interactive Cartographic Explorer interface are similar to the one of the global exploration map).
The central panel allows the user to build a map of objects of interest
· double-clicking on image to develop other images
· right-clicking on image to suppress them
· rolling wheel to zoom in and out
· click on central button on the mouse to display all the documents
· Shift + mouse to zoom in the left menu
Two additional menus at the right and at the left of the graphic map are provided to help the user to navigate in the map:
The right panel gathers functional aspects:
· Some parameters to customize the graph view
· Setting of the active modality.
The left panel is used to help the user in the navigation
· Document image view: displays the picture of the document currently pointed by the mouse (+ shift)
· Document text view: displays the textual metadata of the document currently pointed by the mouse (+ shift)
Fig. 28 Presentation of the Interactive Cartographic Explorer
This tool allows producing Video news html interactive maps. To launch it, the user has to perform a video search and click on a video in the results. Then the NIM tool may be invoked by clicking on the ‘*’ button in the ‘Zoom’ Window.
Four components compose the NIM interface: the video player, the map view, the textual description view and the search widget (see next Fig.).
Video player and graph view are synchronized. The current image of the video player has a red frame in the graph view. The images of the report being played have orange frames. Clicking on the current image controls the play/stop video functions.
Textual descriptions of the current report are displayed on the lower left area of the panel.
Two navigation modes have been implemented. They may be switched in double clicking on the graph view.
· video overview mode (default) allows the user to parse quickly the video just by moving the mouse on the images of the graph view
· Precise display: allows the user to control the video player by clicking on the images of the graph view.
Textual input may be given on the search textual field, segments results of the query are marked by a colour offset.
Moving the mouse on the textual description highlights the corresponding segment in the graph view.
A User’s Profile comprises the properties, the preferences, the role and the access profile corresponding to a specific user. Profiles are the selected mechanism to implement personalization in VITALAS. They provide the capacity to customize the system based on user’s role, preference selection, search behaviour and interactive feedback
The Profile of a user includes four subsets of information: Personal data, Role, Access Profile and Preferences:
· Personal data include username, password, name, surname, company, job position, domain, and contact data. These data will be provided by the user at the Registration process and some of them can be modified by him.
· The Role is assigned by default at registration time and can be later modified by the Administrator.
· Access Profile (access rights to data sources) is assigned by default at registration time and can be later modified by the Administrator.
· Preferences are set from a default profile and can be modified by the user according to his/her preferences on some parameters that control the application style, the search process and the pages’ layout.
The user has access to his/her profile by clicking on the link “My Profile” in the VITALAS top menu. As a result a submenu is displayed providing access to three user profile options:
The user’s account option gathers some personal and contact data provided by the user at the registration process (see next Fig.). It also includes role and access profile.
By clicking on the “Edit” button the user can change some fields, such as password, job position, domain, contact data, etc. whereas other fields, such as role or access profile, cannot be modified.
Fig. 30 My Profile -> My Account
My Preferences section gathers some preferences defined by the user that allow the
adaptation of the system to his specific needs. Preferences have been
classified in two groups (tabs): Preferences for Search and Preferences for
results
In this section the user can define default values to launch search processes (see next Figure). Nevertheless, all these default values can be further modified by the user during the search:
· Search type (basic, extended, advanced)
· Media type (image, video)
· Starting page (welcome page, search page, administration page)
· Topics of interest (fixed keyword entries for which the user can select up to ‘n’ values)
· Collections/Sources to search (all the image datasets and video datasets that have been defined in the user’s access profile; the user can select a subset of them)
· Orientation
· Style (colour, black & white)
· Rights (rights managed, royalty free)
· Video search on (text, audio)
Fig. 31 My Profile -> Preferences for Search
In this section the user can define the default way in which the system presents the results of a search. Those sections/fields that are initially hidden according to users’ preferences can be further displayed on user’s request:
· Results presentation style (mosaic view, cluster view, advanced view)
· Some configuration parameters for each view (size of items per page, sort criteria)
· Metadata fields to be shown in the Image Details page and in the Video Details page
Fig. 32 My Profile -> Preferences for Results
The VITALAS team proposes you the following
tasks to evaluate the system before fulfilling the questionnaire. Of course,
you are free to follow this guide or test the system in any other way you would
prefer.
T1: Find images of athletics (sprinters
on lanes)
T2: Find images of red formula one car
T3: Find images of a soccer player (image
with player from head to foot shooting the ball)
T4: Find images of a boat (context: leisure)
T5: Find images on which Adidas appears
as sponsor (Adidas logo visible)
T6: Find images of finishers of a cycling
race (road bikes)
T7: Find images of traffic jam (no accident)
T8: Find images of red flowers
T9: Find images of the wedding of Kim Clijsters
(the former tennis player) where she's portrayed with her husband (Brian Lynch
- a former basket player)
T10: Find images of images of Pope Benedikt
on his visit to
T11: Find images on which the UNICEF appears
as supporter (logo visible)
T12: Find images of a a bicycle (without
a person on it)
T13: Find images of a demonstration crowd
T14: Find images of one or more apples
(fruit)
T15: Find images of photographer Ahmad
Yusni of the Tsunami catastrophe (26.12.2004)
T16: Find images of Paris Hilton on a red
carpet wearing a colored dress
