TR2015-111

Depth-weighted group-wise principal component analysis for Video foreground/background separation


    •  Tian, D.; Mansour, H.; Vetro, A., "Depth-Weighted Group-Wise Principal Component Analysis for Video Foreground/Background Separation", IEEE International Conference on Image Processing (ICIP), DOI: 10.1109/ICIP.2015.7351400, September 2015, pp. 3230-3234.
      BibTeX Download PDF
      • @inproceedings{Tian2015sep,
      • author = {Tian, D. and Mansour, H. and Vetro, A.},
      • title = {Depth-Weighted Group-Wise Principal Component Analysis for Video Foreground/Background Separation},
      • booktitle = {IEEE International Conference on Image Processing (ICIP)},
      • year = 2015,
      • pages = {3230--3234},
      • month = sep,
      • publisher = {IEEE},
      • doi = {10.1109/ICIP.2015.7351400},
      • url = {http://www.merl.com/publications/TR2015-111}
      • }
  • MERL Contacts:
  • Research Area:

    Digital Video


We propose a depth-weighted group-wise PCA (DG-PCA) approach to separate moving foreground pixels from the background of a video acquired by a moving camera. Our approach utilizes a corresponding depth signal in addition to the video signal. The problem is formulated as a weighted l2,1- norm PCA problem with depth-based group sparsity being introduced. In particularly, dynamic groups are first generated solely based on depth, and then an iterative solution using depth to define the weights in l2,1-norm is developed. In addition, we propose a depth-enhanced homography model for global motion compensation before the DG-PCA method is executed. We demonstrate through experiments on an RGBD dataset the superiority of the proposed DG-PCA approach over conventional robust PCA methods.