VisionGo: towards true interactivity

  • Authors:
  • Yan-Tao Zheng;Shi-Yong Neo;Xiangyu Chen;Tat-Seng Chua

  • Affiliations:
  • National University of Singapore, Singapore;National University of Singapore, Singapore;National University of Singapore, Singapore;National University of Singapore, Singapore

  • Venue:
  • Proceedings of the ACM International Conference on Image and Video Retrieval
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents our interactive video retrieval system VisionGo. To achieve high interactivity, the system is fitted with a mouse stroke labeling interface, in which users can use mouse to draw the bounding box of target object via a simple mouse stroke. To enhance retrieval efficacy, the system leverages a new label class "near positive", besides "positive" and "negative" label classes. This class of shots allows user to indicate shots that are incorrect but may be useful during feedback.