VisualGREP: A Systematic Method to Compare and RetrieveVideo Sequences

  • Authors:

  • Rainer Lienhart;Wolfgang Effelsberg;Ramesh Jain

  • Affiliations:

  • University of Mannheim, Praktische Informatik IV, 68131 Mannheim, Germany&semi/ University of California at San Diego, Visual Computing Lab, La Jolla, CA 92093-0407, USA. lienhart@pi4.informatik.u ...;University of Mannheim, Praktische Informatik IV, 68131 Mannheim, Germany&semi/ International Computer Science Institute, 1947 Center Street, Berkeley, CA 94704-1198, USA. effelsberg@pi4.informati ...;University of California at San Diego, Visual Computing Lab, La Jolla, CA 92093-0407, USA. jain@ece.used.edu

  • Venue:
  • Multimedia Tools and Applications
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we consider the problem of similarity between videosequences. Three basic questions are raised and (partially)answered. Firstly, at what temporal duration can video sequences becompared? The frame, shot, scene and video levels are identified.Secondly, given some image or video feature, what are therequirements on its distance measure and how can it be “easily”transformed into the visual similarity desired by the inquirer?Thirdly, how can video sequences be compared at different levels? Ageneral approach based on either a set or sequence representationwith variable degrees of aggregation is proposed and appliedrecursively over the different levels of temporal resolution. Itallows the inquirer to fully control the importance of temporalordering and duration. The general approach is illustrated byintroducing and discussing some of the many possible image and videofeatures. Promising experimental results are presented.