Motion Detection and Segmentation using Image Mosaics

Authors
Kiran S. Bhat
Mahesh Saptharishi
Pradeep K. Khosla

Abstract
We propose a novel motion segmentation algorithm for extracting foreground objects with a pan-tilt camera. Segmentation is achieved by spatio-temporal filtering of the scene to model the background. Temporal filtering is done by a set of modified AR (Auto-Regressive) filters, which model the background statistics for a particular view of the scene. Backgrounds from different views of the pan-tilt camera are stitched together into a planar mosaic using a real-time image mosaicing strategy. Our algorithms work in real-time, require no user intervention and facilitate high quality video transmission at low bandwidths.

Paper
Download the paper (ICME 2000):

Results
Motion detection using AR Filters:
  • Left is input sequence, right is the estimated background model. avi (7.4MB)
Motion detection with image mosaics:
  • Background image mosaic. avi (0.5MB)
  • Input Sequence. avi (1.5MB)
  • Detection using the mosaic background. avi (0.5MB)