Home . Projects . Research . Publications . Members . Demos
The Fact Extractor tool is used to extract spatio-temporal relations between video objects and store them in the knowledge-base as facts. These facts representing the extracted relations are used to query video data for spatio-temporal conditions. The tool also extracts object appearance relations and object trajectories. A demo facts-base population is available here.
The Object Extractor tool, which has been added to the Fact-Extractor as a module, is used to extract salient objects from video keyframes. It also facilitates the fact-extraction process automating the minimum bounding rectangle (MBR) specification of salient objects. A demo of the Object Extractor module is available here.
The Video Annotator tool is used to annotate video clips for their semantic content and to populate the feature database of the BilVideo system for semantic queries on video data. It provides some facilities for viewing, updating and deleting semantic meta-data that has already been created for video clips and stored in the feature database. A demo of Video-Annotator tool is available here.
The BilVideo system has a Web-based query interface for specifying spatio-temporal and object trajectory queries to the system. The query specification is handled simply by positioning visual components on the drawing canvas. The visual components are minimum bounding rectangles of salient objects for spatio-temporal queries, whereas the visual components for object trajectories are paths of knobs representing the turning points of the object trajectories. A demo of the Web-based Query Interface is available [http://pcvideo.cs.bilkent.edu.tr/|here]. The interface is still under development and the final version of the Web-based Query Interface can be seen here.
This tool is used to query images and video keyframes on color and shape features. The tool uses similar techniques as in Object Extractor to extract objects from the image data, and constructs histogram-based model for representing the color and shape content. Weights can be assigned to shape and color features during querying, and query results are presented to the user in descending order of similarity.
This tool is used to query images and video keyframes on color and shape features. The tool provides three types of color and texture querying on images: whole image, semi-automatically specified region, automatically extracted region. This nature of the tool is used to query wide range of image domains. Weights can be assigned to shape and color features during querying, and query results are presented to the user in descending order of similarity.
The KiMPA tool is used to facilitate the polygonal representation of extracted objects. The system requires polygonal object data, which can be a polygon having large number of vertices. However, processing these huge representations may lead to some inefficiencies, thus an approximation is required. With this tool, a polygonal object can be simplified to a smaller n-sided polygon, where n is a user-specified parameter.