4. Interactivity in Video Database Applications

4. Interactivity in Video Database Applications

The two systems developed at SCR are both interactive video database applications that use a video-driven composition technique, similar to the one described by Auffret et al. [17]. The content model underlying this video-driven approach is illustrated in Figure 16.3.

Figure 16.3: The Interactive Video Content Model

4.1 The Video-Driven Composition Technique

The content model being used by the video-driven technique is comprised of the following elements (see also [18]):

• Video sequence element (clip). The video consists of a sequence of video clips called sequence elements. Each such sequence element is a contiguous sequence of frames from a video recording. A sequence element may contain multiple shots if the corresponding video recording contains multiple shots in sequence. Each sequence element keeps the reference to the video recording in which it is defined. Hence the model supports the concatenation of clips from potentially different video recordings. It should be noted that this simple form for composition is not considered as a replacement for a video authoring tool. The objective is for the model to define an assembly of pre-edited video sequences that allows for easy repurposing of such video content.

• Video section. Contiguous sequence elements are grouped into sections. A section typically represents a meaningful unit of content - e.g., a news story. Sections are used as targets for direct access, i.e., the user may use the table of contents (see below) to start the playback at the very beginning of any given section.

• Video hyperlink. There may be hotspots within the video. Hotspots have a spatial and a temporal range. Two hyperlinks are shown in the example model in Figure 16.3. The temporal range is illustrated by the grey shading. The broken arrows indicate that the target object for the hyperlink will be loaded in one of the application's content frames if the end-user activates the corresponding hotspot.

• Synchronization point. Synchronization points are temporal locations within the video that defines some transition or change of content in one of the application's content frames. Synchronization points are shown as small dots on the figure. Synchronizations points are useful, for instance, in cases where the interactive video was recorded during a lecture or a business presentation and where the synchronizations points are used to synchronize the "playback" of the slides being used during the presentations with the video playback.

• Table of contents. The content of each interactive video is listed in a table of contents. Each entry in the table of content is composed of the following fields:

  • Section title. The title of the section, which will appear in the table of content.

  • Section abstract. A textual description of the content of the section, which will also be shown in the table of content.

  • Section URL. A URL that identifies the beginning of the section within the video production. The arrows in Figure 16.3 pointing from the entries in the table of content to the interactive video section indicate the presence of section URLs.

4.2 Interactive Video Tools

The first part of this section discussed the data stored in the video meta-database for generating interactive video applications. Tools are needed for creating the presentations from the data stored in the database and for creating and managing these data. The remaining part of this section discusses such tools.

4.2.1 Presentation Generation

Generating the presentation is a process consisting of three logical steps, as shown in Figure 16.4. The first step consists of selecting what parts of the interactive video will be delivered. The personalization process used in HotStreams™, for instance, selects the parts of the interactive video that fits the end-user's interest profile. The ultimate goal of the process is to generate scripts that will drive the interactive video presentation. Unfortunately, there exist several script languages that might be used to drive an interactive video production. Hence, it is beneficial to generate an internal, language independent representation of the production first. Support for multiple script languages can then easily be incorporated by implementing a script generator for each of the languages that convert the video from the internal representation to the specific script language.

Figure 16.4: The Presentation Generation Process

4.2.2 Content Assembly Tools

HotStreams™ offers a web-based tool for managing the contents of the meta-database [18]. The tool is implemented as a Java applet that runs inside a regular web browser. The tool consists of a number of panels where each panel implements a group of closely related management functions. The video panel shown in Figure 16.5, for instance, provides the means needed to create and manage interactive videos, their composition, and the text fields that constitute the table of contents. Similarly, the hyperlink panel provides the means needed to create hyperlinks and define their appearance and destination. These panels also provide means to generate meta-data that will be used to create personalized content.

Figure 16.5: The HotStreams™ Content Management Tool