`Proc. of SPIE-IS&T Electronic Imaging, SPIE Vol. 6074, 60740O, © 2006 SPIE-IS&T · 0277-786X/06/$15
`
`SPIE-IS&T/ Vol. 6074 60740O-1
`
`A Mobile Phone-based Context-aware Video Management
`Application
`
`Janne Lahti, Marko Palola, Jari Korva, Utz Westermann*, Kostas Pentikousis and Paavo Pietarila
`
`VTT Technical Research Centre of Finland, FI-90571 Oulu, Finland
`e-mail: firstname.lastname@vtt.fi
`* University of California, Irvine, CA 92697-3425, USA
`e-mail: westermann@acm.org
`
`ABSTRACT
`We present a video management system comprising a video server and a mobile camera-phone application called
`MobiCon, which allows users to capture videos, annotate them with metadata, specify digital rights management (DRM)
`settings, upload the videos over the cellular network, and share them with others. Once stored in the video server, users
`can then search their personal video collection via a web interface, and watch the video clips using a wide range of
`terminals. We describe the MobiCon architecture, compare it with related work, provide an overview of the video server,
`and illustrate a typical user scenario from the point of capture to video sharing and video searching. Our work takes
`steps forward in advancing the mobile camera-phone from a video playback device to a video production tool. We
`summarize field trial results conducted in the area of Oulu, Finland, which demonstrate that users can master the
`application quickly, but are unwilling to perform extensive manual annotations. Based on the user trial results and our
`own experience, we present future development directions for MobiCon, in particular, and the video management
`architecture, in general.
`
`Keywords: mobile camera phone application, mobile video management, video metadata and digital rights management
`
`1. INTRODUCTION
`Mobile phone manufacturers are increasingly adding new models with multimedia support and most modern medium- to
`high-end cell phones come with an integrated audio/video player, a camera to capture still and moving pictures, and
`some media editing software. The “coolness factor” fuels the popularity of mobile camera phones (MCP) and increases
`the volume of user-created media content. MCPs can record videos of up to several minutes, depending on the amount
`of memory available. Videos cannot be reasonably stored permanently on the mobile device due to the limited memory
`capacity available. Thus, users are in need of services that allow them to store their videos somewhere else, also because
`users want to create collections of their clips and share them with friends and relatives. Some will even opt for making
`part of their digital content available from a web site, or add it to their blog entries. Although none of these is news,
`nevertheless, there is very little automation in this process, and mobile applications for video management are not as
`commonplace as one might think despite the fact that MCPs have been on the market for quite some time.
`The main problem of mobile content management is two-fold: how to automate permanent video clip storage, and
`how to do so in a way that is user-friendly, allows for easy clip lookups, and enables the user to share videos with others.
`Even though there are standardized ways to share videos over the cellular network, including the Multimedia Messaging
`Service (MMS) [1] and the upcoming IP Multimedia Subsystem (IMS) [2, 3], they must all be supported by the mobile
`operator, require infrastructure expenditures, and allow mainly for point-to-point video sharing. Moreover, a common
`characteristic of these services is that they do not include a solution for video management, and do not take full
`advantage of the context information available at the point of video capture. Our development effort takes a different
`approach and attempts to (a) enable the MCP to serve as an valuable video production tool in addition to being a video
`consumption channel, and (b) permit users to share videos with others irrespective of the level of network operator
`support for multimedia services.
`Of course, one can argue that an MCP is not the ideal device for video management because of the limited CPU,
`memory and input/output capabilities, battery power consumptions considerations, and the diverse video formats and
`
`Page 1 of 12
`
`
`
`I G
`
`F
`
`SPIE-IS&T/ Vol. 6074 60740O-2
`
`other technologies supported by the operating system. On the other hand, a MCP application can assist users to describe
`a video by associating it with annotations about, for example, where and when a video clip was shot, and who should be
`able to access it. The MCP application may have at its disposal different kinds of context information including sensors,
`a GPS device, a calendar service, and other Internet resources. All these are valuable sources for metadata that can be
`used to characterize the digital object and possibly uniquely identify it. Metadata plays a critical role in managing video
`content to the extent that Sarvas et al. [4] argue that it is not possible to manage media content effectively without them.
`The value of the metadata is illustrated in the following section, which presents a typical scenario where the user
`captures a video clip, associates it with metadata, and later searches for this particular video within her collection using
`an Internet-connected desktop PC. This paper details a solution to the problems related to user-created media content
`and annotation, and addresses some issues pertaining to sharing and storing videos using MCPs, and the challenges
`presented to the mobile video management system. The rest of this paper is organized as follows. Sections 3 and 4
`discuss the implementation objectives of such video management system, and its architecture, respectively. Section 5
`summarizes the results of a real user field trial evaluation, and Section 6 presents our current work in progress along with
`some future directions, which address certain shortcomings found in the first generation of the system. Finally, Section 7
`reviews related work, and Section 8 concludes the paper.
`
`2. A TYPICAL USER SCENARIO
`We exemplify the use of the mobile video management system with a scenario involving a tourist in a sightseeing tour.
`The aim in this section is to provide an overall understanding of how the system works, essentially a high-level user-
`centered view of the requirements, before presenting, in the following section, the main objectives.
`Figure 1 portrays Alice while on vacation in Oulu, Finland. Alice spends some time walking around the city center
`area, visiting several attractions, when she notices the Oulu cathedral clock tower from a distance. After approaching the
`church, she uses her MCP to record a short video of the cathedral, its clock tower, and the surrounding area. Alice uses
`MobiCon, an MCP application, which allows her not only to capture the video, but also to upload it to a server and notify
`her friends about recent findings during her tour of Oulu.
`
`
`
`
`Figure 1: Alice records a video of the Oulu cathedral (left), annotates it with the term “Object > Buildings” (center), and later is able to search for it
`using the web interface of the Candela video management server (right).
`
`After recording the video, Alice annotates it by selecting the prefedefined concepts “Holiday” and “'Buildings” from
`the application metadata menu, and enters two keywords (“Church” and “Oulu”) to describe the video clip more
`accurately. Alice saves the clip locally on the mobile phone memory card. Shortly afterwards, Alice decides to share the
`video with Bob, a friend of hers who is very interested in church architecture. With MobiCon, this is a simple process:
`Alice selects the video clip using a menu, chooses Bob from her contact list, and grants him the rights to watch the clip.
`She can subsequently upload the clip to the video server and MobiCon will automatically send a text message using the
`Short Message Service (SMS) to Bob with information on how to access the video. After receiving the text message,
`Bob can watch the video by opening its URL straight from his mobile phone.
`After returning from her holidays, Alice can login to the video server using a web interface and access her large clip
`collection. Mobile videos tend to be short in length and relatively “focused” on a single theme. Unfortunately, Alice
`quickly finds herself having to deal with hundreds of videos with not-so-descriptive file names. Trying to locate the
`
`Page 2 of 12
`
`
`
`SPIE-IS&T/ Vol. 6074 60740O-3
`
`“church video” in the entire collection can quickly become tiresome and this is where video metadata annotation proves
`helpful. Although Alice is not likely to remember the exact time or date that the video was captured or uploaded to the
`video server, she does remember where she was and what the main theme of the video was. Searching for “Oulu” limits
`the set of possible videos and searching for “church” narrows the size of the search result set to a handful of clips. Once
`located, Alice can watch the movie clip delivered by the video management server in the most appropriate format for her
`platform.
`
`3. IMPLEMENTATION OBJECTIVES
` A mobile video management application has five main functions: video recording, metadata annotation, video storage,
`video sharing, and locating video clips in a collection. Any MCP application, such as MobiCon, should be robust and
`rich in functionality, yet easy to use and engaging, despite the restrictions imposed by the small display size and minimal
`keyboard. Moreover, application developers must pay attention to the way resources are used: network traffic should be
`minimized, battery power should be conserved when possible, and CPU and memory ought to be utilized with frugality.
`These restrictions come on
`top of
`the classic mobile phone application development nightmares (device
`incompatibilities, network application debugging, immature SDKs, and different operating system versions with
`undocumented bugs) making the development of an application like MobiCon challenging.
`Video recording, the first function, is relatively straightforward to implement with vendor provided SDKs. However,
`the application should be robust during this phase and capable of handling critical events (including incoming phone
`calls and text messages). As illustrated in the previous section, video metadata annotation is necessary for searching
`stored clips in an efficient manner and is a central part of the design of MobiCon. We can identify three main issues that
`need to be addressed: (a) when shall the application collect the metadata, (b) which types of metadata should be stored so
`that locating the video will be easier later, and (c) how can the application acquire all this information in an elegant and
`robust manner. Practice indicates that, at least for home videos, the best time to annotate video clips is right after
`capturing them. However, at this point it is difficult to predict which information will be most valuable in locating the
`video later on. For example, for a resident of Oulu using MobiCon frequently while in town, the term “Oulu” will not be
`a good choice in order to discriminate between a large set of videos. For Alice, though, the same term will allow her to
`easily select all videos from her last visit to Northern Finland. Thus, one can argue that the best choice is to collect as
`much information as possible, and preferably do so in an automated way. On the other hand, this may lead to generating
`mostly low-level data or content features, such as exact dates and color histograms, which are neither easily recollected
`by most humans nor convey a direct meaning to them. Therefore, the application must allow the user to provide
`additional high-level information, and assist him in this by making the process as easy as possible.
`The limited phone resources make video storage and video sharing particularly interesting problems to address.
`Permanent storage cannot be provided by the phone in a scalable manner: compared to other kinds of mass storage
`devices, such as hard drives, memory cards for mobile phones are still expensive and very limited in capacity. Users
`typically end up transferring their videos to their PC and then share them with others. This, of course, is not a truly
`mobile video management solution. An MCP application should use the network to store the videos and allow others to
`view them in an asynchronous manner. For video sharing, there are no universally supported media formats, and the
`device capabilities and the capacity of the access networks vary greatly. Because of this, the original video ought to be
`provided in several alternative formats using different encoding parameters. However, video transcoding is a
`computationally demanding process which cannot be performed in real time. Thus, alternative versions should be
`generated before the video can be retrieved, typically soon after a clip is uploaded. Furthermore, an important aspect in
`video sharing is the ability to control how the receivers use the shared video clips and limit redistribution, if necessary.
`One way of handling the video storage and sharing is to place the videos on a server. This way the MCP application
`needs to upload the video only once while remaining able to share it with an arbitrary number of recipients. The server
`can also take care of all security and computational intensive aspects of video storage, and enable the user to search for
`video clips later. In short, by making an MCP application directly pluggable to a searchable repository that takes
`advantage of the captured metadata leads to avoiding time-consuming and inefficient lookups of video clips based on
`filenames and file dates alone.
`To sum up these requirements, MobiCon should be able to capture video clips using the internal mobile phone
`camera, assist the user to annotate each clip with metadata, store the clip to a remote video management server, permit
`him to share video clips with others and, last but not least, enable him to search large collections of video clips using a
`mobile phone, PC, or any other device with web access. Users should maintain full control of the entire video man-
`agement chain, from source to destination, by tapping on a standards-compliant digital rights management (DRM)
`framework.
`
`Page 3 of 12
`
`
`
`qsbpAG MGP fli äGUGLS{OL
`
`oqncou suq
`bLG2Gufspou f!GL
`
`OUf0I0ä1
`WSUSäGL
`
`VUUOfSf!OU
`
`2GLA!CG
`
`!qGo nbosq ä9{GM&
`2fLGSW!Uä 2GLAGL
`!qGo wsusäGL
`
`GUä!UG
`
`!qGo wsusäGwGuf suq
`bLocG22iuä {!GL
`
`2{OLSäG {!GL
`
`SPIE-IS&T/ Vol. 6074 60740O-4
`
`4. SYSTEM ARCHITECTURE
`The VTT Candela system, named after the European ITEA project CANDELA (Content Analysis, Networked Delivery
`and Architectures) was developed as a solution for general video management. It includes tools for video creation,
`analysis, annotation, storage, search and delivery phases. The Candela system was originally developed for the personal
`home video domain but, due to its modular and tiered architecture, it can be used as a basis for different video
`management applications by reusing some of the generic components and adding domain-specific ones. An important
`part of the entire architecture is the MCP application MobiCon, which can be used both on a standalone basis and as part
`of the entire system, as shown in Figure 2. Before delving in the details of the MobiCon architecture, we briefly
`introduce the Candela system; more details about the Candela architecture are given in [5, 6].
`
`
`Figure 2: Candela and MobiCon system architecture.
`
`
`
`4.1. Candela
`Users can add video content to the Candela system produced by several different types of sources. The traditional
`method for home video content is to digitize analog content from a camcorder, or to use a digital one and upload the
`content to the system. Besides the home video domain, the Candela system can be a part of an enterprise-centered
`application such as, for example, an automatic security system handling the video content generated by a set of
`surveillance cameras. In the domain of personal mobile multimedia, the content is created in a mobile phone with video
`capture capabilities. In all application domains, the content is annotated automatically or semi-automatically, utilizing
`content analysis methods and input from sensors such as GPS positioning devices or context information which is
`available for example in a user’s calendar. This information is captured in MPEG-7 metadata descriptions [7], which are
`stored along with the actual video data to support searching and managing video storage.
`Candela uses the commercial Solid FlowEngine relational database system which provides scalability from a single
`in-memory database on a miniature embedded device to complex distributed and duplicated fault tolerant settings, which
`allows us to deploy more advanced configurations in the future. When storing, MPEG-7 descriptions are mapped to a
`relational schema so that SQL can be used for querying. In order to provide better results, the Candela query engine
`broadens user queries by suggesting additional search terms, which are, based on a domain ontology, closely related to
`the ones specified by the user or more descriptive. On the other hand, user profiles are used to restrict the amount of
`found matches to those that are the most relevant to the user.
`Candela supports a very broad set of end user terminals ranging from cellular phones to high-end desktop computers.
`In order to provide a user friendly experience, the web-based interface dialogs are generated dynamically by using open
`source Apache Cocoon framework for XML transformations. This allows us to customize the amount of information
`presented to the user and the dialog between the user and the system to the capabilities of the used access device. Once
`the user has found an interesting video, a streaming video player is launched. This is either one of the off-the-self video
`players or Candela video browser which shows visualized metadata in addition to the video itself. The visualization
`allows the user to navigate through the video and find the relevant parts of it easily [5].
`Providing optimal video quality for the end-user given the diversity of the source material, differences in user
`terminal capabilities and characteristics of networks, especially in the mobile domain, is a challenging task. MPEG-4
`
`Page 4 of 12
`
`
`
`SPIE-IS&T/ Vol. 6074 60740O-5
`
`standard offers some scalable video coding solutions, where the changing network capabilities for video delivery can be
`taken into account in real-time by inserting enhancement layers to the video stream in case of more available bandwidth.
`However, it was concluded in the project that the state-of-the art in scalable video coding does not offer reasonable
`quality as compared to the non-scalable stream at the same bitrate [8].
`If the media source is a mobile phone with a low resolution camera, the need for transcoding is not as obvious as for
`more bandwidth consuming content, but overall we want media to be accessible across platforms and at the moment
`there are no uniformly supported video formats. Thus the solution was, at the expense of storage, to transcode the
`material to a representative set of formats and bitrates and develop a content negotiation plug-in for Helix streaming
`server in order to choose from those.
`
`4.2. MobiCon
`The MobiCon client-server architecture is shown in Figure 3. MobiCon consists of two different software components:
`the UploadClient, which is a mobile Java (J2ME) application running on a mobile phone and UploadGateway, which is
`implemented as a Java servlet in the Candela server. Architecture is based on the need to provide wireless access over a
`mobile phone network to enable storing video clips on the server where it is also possible to run more computation-
`intensive operations such as video transcoding.
`
`
`Mobile Phone
`
`Mobile Network
`
`Server
`
`User
`Interface
`
`UIManager
`
`Videodata
`
`UserManager
`
`ConnectionManager
`
`(cid:127) Video data
`(cid:127) Metadata
`(cid:127) DRM settings
`(cid:127) User
`authentication
`
`LoginManager
`Servlet
`
`VideoManager
`Servlet
`
`SMSNotifier
`
`(cid:127) Access key
`
`(cid:127) User info
`
`DRMDelivery
`Servlet
`
`Userdata
`
`Metadata
`
`Videodata
`
`Candela
`Interface
`
`Upload Client
`
`Upload Gateway
`
`URL-link to video clip
`
`DRM protected video clip
`
`Figure 3: High-level description of MobiCon.
`
`
`
`MobiCon naturally needs to be easily installed without any extra tools or additional instructions. The server allows
`distribution of MobiCon application easily to mobile phone users by using Over-The-Air (OTA) specification from the
`Open Mobile Alliance, which enables mobile applications to be downloaded and installed over the cellular network.
`After installation, the user is authenticated by the server using a username and password to log on the Candela system.
`The username and password are transferred to the UploadGateway and as a reply to successful authentication user profile
`information is transferred back to the UploadClient where UserManager stores user information (name, address, etc.),
`which are also used as metadata of captured video clips. The UploadClient no longer asks for username and password
`after the first time: for user convenience, it is assumed that the user stays the same after the first. Mobile phones are
`personal devices which are rarely lent to other people or left unguarded, the decision was made that the risk is not too
`large compared to the benefits achieved.
`The UIManager is a controller component which is loaded first when the application is started. The UIManager
`coordinates the video capture using the mobile phone's camera, the saving of the video data to the Java Record Store
`system, and the sending of video sharing SMS messages to the other users. UIManager also provides user interfaces that
`are presented in the next Section. The ConnectionManager handles the connection between the UploadClient and
`UploadGateway providing data transfer using HTTP-protocol over the packet networks such as GPRS/EDGE/WCDMA.
`ConnectionManager delivers the captured video data, its metadata, user name, and DRM options to the Upload Gateway.
`If the user had chosen DRM protection for content sharing, the UploadGateway reply contains the identification keys to
`
`Page 5 of 12
`
`
`
`anlanne
`
`1. Capture video clip
`
`actianor tuSpr tiPtiomn
`
`Sto
`
`1 -
`
`optiom
`
`2 LOgging 0..
`
`otr
`
`Oaflfle.Z9.11.OS_ZO:1U ;60k to main
`
`Options
`
`lanne.ZY.11.OSZO:110pti005
`
`Close
`
`options
`
`OCR
`
`SPIE-IS&T/ Vol. 6074 60740O-6
`
`be shared with another user. The ConnectionManager creates a URL containing the received identification keys and
`returns them to the UIManager, which sends them to the other users via the SMSNotifier.
`The UploadGateway serves multiple MobiCon users and provides access to the Candela system. Figure 3 presents an
`architectural overview of the UploadGateway. The LoginManager servlet handles all the tasks needed for user
`authentication on the server side. The VideoManager servlet takes care of all the functionalities receiving video data
`from the UploadClient to the UploadGateway. The received video and metadata descriptions are stored temporarily, the
`video clip is transcoded, a key frame picture is extracted from the video, and metadata is finally formatted to the MPEG-
`7 XML format. All data including video clips, keyframes, and MPEG-7 are added to the database via Candela Interface.
`The DRMDelivery servlet handles DRM video packaging and sharing. It is based on the Nokia Content Publishing
`Toolkit (NCPT) that allows creation of a DRM-protected video file package called Download Descriptor (DD), which
`holds the author, video name, size, and description. The DRMDelivery servlet creates a random number for DRM
`package name, which is delivered to other users in a URL address pointing to DRMDelivery servlet by SMS Notifier.
`The receiver of the shared video file needs only to open the URL link and to connect to DRMDelivery servlet using web
`browser in mobile phone and access the video clip if the DRM system in the mobile phone allows the access. The
`receiver can also see who the sender was and what kind of content is shared.
`The MobiCon shared videos can be played directly via RealVideo player on mobile phone. With the DRM delivery
`channel a user can share video clips with other users who have a mobile phone supporting the OMA DRM standard and
`have a video player. DRM is an umbrella term for a set of technologies developed to protect against unauthorized
`copying and distribution of copyrighted material [9]. MobiCon is using DRM v1, although all mobile phones might not
`handle DRM protected content in standardized style [10].
`
`4.3. MobiCon Interface Flow Diagrams
`This section presents the MobiCon functionality from the user perspective with a walkthrough of typical usage scenarios.
`The user authentication and video capture is shown in Figure 4. The editing/uploading video clip is presented in Figure
`5, which is expanded on Figure 6 and Figure 7 showing metadata annotation and DRM parameterization flows. Each
`figure comprises two parts: on the left side an Interface Flow Diagram [11] is given, while the right side presents the
`actual MCP screenshots.
`User authentication and video capturing (Figure 4) starts with entering username and password for authentication
`with the server (Screenshots 1-2). If the username and password are valid, MobiCon receives the user’s personal profile
`information and stores it in the phone memory. Then, MobiCon's main screen is displayed (Screenshot 3), where the
`user can choose to view and edit personal information, to load video clips, or to capture a new clip (Screenshot 4). A
`new video clip is captured in Capture Screen using Mobile Media API and it is recorded according to 3GPP specification
`using AMR coding for audio and H.263 at 176x144 pixels size at 15 frames per second for video. After a video is
`captured or loaded (Screenshot 5), the user is presented the Edit Screen (Screenshot 6) that ends the user authentication
`and video capturing scene.
`
`
`
`
`
`
`Figure 4: User authentication and video capturing: UI Flow Diagram (left) and UI screenshots (right).
`
` Figure 5 presents the editing and uploading phase starting from the Edit Screen (Screenshot 1), which displays
`different editing options for the video clip. The user can choose to view the video clip in the Video View Screen
`(Screenshot 2), enter metadata annotations in the Metadata Screen (Screenshot 3), upload video clip (Screenshots 4, 5
`and 6) either through the Access Screen or the DRM Screen, save, and capture a new video clip.
`
`Page 6 of 12
`
`
`
`ntate clip
`Ireetewt
`
`4
`
`5. Cap ture new clip
`t1 ,,cktn main
`oution
`
`ciroôfion5
`
`0
`
`6,.iding widen data
`VdlIfl[Idt
`
`5
`
`OUtiD[15
`
`Dtofrons
`
`Humans
`Octivities
`Objects
`Environment
`o!,din
`3
`
`H nimoy
`Sucial_life
`Everyday_life
`Work_activities
`Stody
`4 ts
`
`snip
`Party ____________
`Clob
`O estantan t
`Hirthday
`
`Oution5
`
`Duck
`
`SPIE-IS&T/ Vol. 6074 60740O-7
`
`
`Figure 5: Editing and uploading the captured/loaded video clip: UI Flow Diagram (left) and UI screenshots (right).
`
`
`
`In the Edit Screen, the user may choose to upload the clip to the Candela server. The Upload Screen (Screenshot 4)
`prompts the user to choose the Candela or the DRM delivery channel or both. If the user chooses Candela delivery
`channel the video clip is uploaded to Candela server over mobile phone network and added to the Candela video
`database, from where the user can later search and access the video clip. A user must also choose whether the video clip
`is only for personal use or whether other Candela users can also find the clip via search. This is selected in the Access
`Screen (Screenshot 5). Then the video clip is uploaded via the Upload Gateway to the user video collection on the
`Candela server (Screenshot 6). In the server the video clip is handed over to the Video Manager Servlet, which
`transcodes the video clip into different formats and bit rates in order to provide a scalable service quality for different
`devices and network connections. Currently, the Video Manager Servlet prepares Real Video, H.264, and H.263
`encodings for delivering the captured video content to mobile devices and MPEG-4 file format for desktop computers.
`In the future, scalable video codecs will remove the need of transcoding [12]. The DRM delivery channel (Screenshot
`4), to prepare video clip for DRM protected sharing, is presented in details in Figure 7. The functionalities under the
`Metadata screen (Screenshot 3) are presented in Figure 6, which illustrates the flow diagram for the metadata annotation
`phase.
`
`
`
`
`
`Figure 6: Giving metadata annotations: UI Flow Diagram (left) and UI screenshots (right).
`
`
`
`The clip annotation starts from the Metadata Selection Screen (Screenshot 1) where the user has options to annotate
`clip using terms in a predefined domain ontology stored in the user's personal profile (Screenshots 3-5), add text
`descriptions by typing the nine-button keypad (Screenshot 2), and edit settings to update personal profile information,
`which is automatically added to the video clip metadata. When annotating via the Ontology Screen, the domain ontology
`can be accessed along a three-level hierarchical structure: the main level view (3) contains top-level concepts, which are
`then expanded in the sub-level menus (4, 5). All three levels are recorded to the clip. MobiCon also gathers context data
`that is available at the time of capture on the mobile phone such as user name, capture time, and duration of the clip.
`Additionally, MobiCon is able to connect to a GPS receiver supporting the NMEA protocol via Bluetooth. GPS provides
`
`Page 7 of 12
`
`
`
`Pawt
`
`1. CANDELA
`
`6630
`
`1. F orwor d—Loc
`
`ath fi wer
`r
`
`44
`
`7
`
`Seodiog video data
`valinnat
`
`Options
`
`B s&ect
`
`Options
`
`Back
`
`.trt.y Ovetumberoftttesthec'pcan
`be pleated!.
`
`Ovastarac
`i. set Nombet of vi 29.11.2005
`2. set viewiot day
`3. 000e
`4 One WeeksiJnlimi________________
`5. One Monthaunlim
`
`Ova stop date (dd.mta.pp
`
`OS BWOS
`
`OffôRS
`
`Dun
`
`SPIE-IS&T/ Vol. 6074 60740O-8
`
`longitude and latitude coordinates and a timestamp for the metadata. All the video clip metadata is uploaded with the
`video clip. In the final MPEG-7 file the context data includes the creator's name, region and country, date and time of
`day, GPS information, and length of the video clip as well as terms given by user and from the ontology tree are
`embedded in free text annotation elements the MPEG-7 provides.
`Figure 7 presents the flow diagram for the DRM protected sharing phase. The starting point is the Delivery Method
`Screen (Screenshot 1) where the user adds the DRM option. The user is prompted by possible recipients from the mobile
`phone's contacts list (Screenshot 2). After selecting a receiver, the user can select between two DRM protection methods:
`Forward-Lock and Combined-Delivery (Screenshot 3). The Forward-Lock method prevents the receiver from resending
`the shared video clip from their mobile phones and it does not require additional configuration. The Combined-Delivery
`method (Screenshot 4) is used to specify access rights for the shared video clip. With the Combined-Delivery method,
`the user can set the dates the receiver can watch the video and how many times the video can be played by the receiver
`(Screenshot 5). After completing the Rights Selection Screen, the UploadClient sends the video, metadata and DRM
`information to the UploadGateway, which creates a DRM package at the server side and returns one identification key
`for each receiver. As soon as Upload Client receives the key(s) it generates and sends SMS messages to the receivers,
`who can download and play the video clip without MobiCon client software.
`
`
`
`
`
`Figure 7: Sharing video DRM protected video clip: UI Flow Diagram (left) and UI screenshots (right).
`
`
`
`5. FIELD TRIAL EVALUATION
`Before initiating the field trial, MobiCon was tested extensively by a small group of developers/researchers from the
`Wireless Services group at VTT. Several issues were identified and fixed in the process, and we felt that both the
`application and the underlying video management service were ready for a field trial with real users. Due to space
`limitations, we wi