A multi-level abstraction and modeling in video databases

Abstract

In this paper, we propose a multi-level abstraction mechanism for capturing the spatial and temporal semantics associated with various objects in an input image or in a sequence of video frames. This abstraction can manifest itself effectively in conceptualizing events and views in multimedia data as perceived by individual users. The objective is to provide an efficient mechanism for handling content-based queries, with the minimum amount of processing performed on raw data during query evaluation. We introduce a multilevel architecture for video data management at different levels of abstraction. The architecture facilitates a multi-level indexing/searching mechanism. At the finest level of granularity, video data can be indexed based on mere appearance of objects and faces. For management of information at higher levels of abstractions, an object-oriented paradigm is proposed which is capable of supporting domain specific views.

Application of Semistructured Data Model to the Implementation of Semantic Content-Based Video Retrieval System

“…Objects can be linked together by the means of explicit relation names. A multilevel abstraction mechanism for capturing the spatial and temporal semantics associated with the various objects in video frames is proposed in [21]. At the finest level of granularity, video data can be indexed on mere appearances of objects and faces.…”

COSMOS-7: Video-Oriented MPEG-7 Scheme for Modelling and Filtering of Semantic Content

“…The key concern here is that of access to digital video resources by users. For all application domains, it is simply impractical for users to spend inordinate amounts of time sifting through an entire digital video resource to find the few items of content that are relevant to them [2][3][4][5][6]. For example, consider that the user wishes to locate video segments within a digital video resource that feature a certain object or depict a certain event.…”

An MPEG-7 scheme for semantic content modelling and filtering of digital video

“…This is crucial with regards content management of large digital video resources since it is impractical for a user to browse through numerous and lengthy video segments, many of which they are likely to not have seen before, to track down content of interest (Day et al 1999;Al-Safadi and Getta 2001;Löffler et al 2002;Jaimes et al 2004). For example, consider that the user wishes to locate segments within a digital video resource where a robbery was taking place and a white van was positioned in front of a bank; or consider that the user wishes to known which events within the video resource feature white vans.…”