Rejection-based Multipath Reconstruction for Background Initialization in Video Sequences with Stationary Objects




This page provides the the generated ground-truth data of the proposed video sequences for the task of Background Initialization. Additionally, the developed software to compute a background image from a video sequence is provided. This material has been used in our submitted paper to Computer Vision and Image Understanding. Moreover, the SBMI2015 dataset [1] has also been used for evaluation.

Contact Information
Diego Ortego -
show email


Background initialization consists in extracting a foreground-free image from a set of training frames. Moving and stationary objects may affect the background visibility, thus invalidating the assumption of many related literature where background is the temporal dominant data. In this paper, we present a temporal-spatial block-level approach for background initialization in video to cope with moving and stationary objects. First, a Temporal Analysis module obtains a compact representation of the training data by motion filtering and dimensionality reduction. Then, a threshold-free hierarchical clustering determines a set of candidates to represent the background for each spatial location (block). Second, a Spatial Analysis module iteratively reconstructs the background using these candidates. For each spatial location, multiple reconstruction hypotheses (paths) are explored to obtain its neighboring locations by enforcing inter-block similarities and intra-block homogeneity constraints in terms of color discontinuity, color dissimilarity and variability. The experimental results show that the proposed approach outperforms the related state-of-the-art over challenging video sequences in presence of moving and stationary objects.


Ground-truth is available here

Results in the proposed dataset and in SBMI2015 dataset are available (proposed and top-ranked algorithms from the literature) here

Software is available here

Currently, software download is disabled as the paper is under review.


[1] L. Maddalena, A. Petrosino, "Towards Benchmarking Scene Background Initialization" in V. Murino et al. (eds), New Trends in Image Analysis and Processing -- ICIAP 2015 Workshops, Lecture Notes in Computer Science, vol. 9281, pp. 469–476, 2015.

© Video Processing & Understanding Lab (