Experimental Results

Results on the Wallflower Dataset

We compared several algorithms in the Wallflower dataset. Figure 1 shows the qualitative results. Table 1 shows the corresponding quantitative results

Figure 1: From top to bottom: original image, ground truth, PCA, RSL, PCP-EALM,PCP-IALM, PCP-LADM, PCP-LSADM, BPCP-IALM. From left to right: MO (985), TD (1850), LS (1865), WT (247), C (251),B (2832), FA (449).

Table 1: F-measure on the Wallflower dataset

For more details, please see :

C. Guyon, T. Bouwmans, E. Zahzah, “Robust Principal Component Analysis for Background Subtraction: Systematic Evaluation and Comparative Analysis”, INTECH, Principal Component Analysis, Book 1, Chapter 12, page 223-238, March 2012.[pdf]

Results on the Background Models Challenge Dataset

We compared several algorithms in the BMC 2012 dataset. Table 2 and Table 3 show the quantitative results.

Table 2: Evaluation results using the synthetic videos for evaluation phase.

Table 3: Evaluation results using the real videos for evaluation phase

For more details, please see :

T. Bouwmans, E. Zahzah, “Robust PCA via Principal Component Pursuit: A Review for a Comparative Evaluation in Video Surveillance”, Special Issue on Background Models Challenge, Computer Vision and Image Understanding, CVIU 2014, Volume 122, pages 22–34, May 2014. [pdf]