Salient video object detection using a virtual border and guided filter

Qiong Wang (1, 2) , Lu Zhang (2) , Wenbin Zou (3) , Kidiyo Kpalma (2)

1 Zhejiang University of Technology
2 IETR - Institut d'Électronique et des Technologies du numéRique
3 Shenzhen Univerisity [Shenzhen]

In this paper, we present a novel method for salient object detection in videos. Salient object detection methods based on background prior may miss salient region when the salient object touches the frame borders. To solve this problem, we propose to detect the whole salient object via the adjunction of virtual borders. A guided filter is then applied on the temporal output to integrate the spatial edge information for a better detection of the salient object edges. At last, a global spatio-temporal saliency map is obtained by combining the spatial saliency map and the temporal saliency map together according to the entropy. The proposed method is assessed on three popular datasets (Fukuchi, FBMS and VOS) and compared to several state-of-the-art methods. The experimental results show that the proposed approach outperforms the tested methods. (C) 2019 Elsevier Ltd. All rights reserved.

Mots clés: Video salient object detection; Distance transform; Guided filter; Global motion; Global motion * Corresponding author

Domaines: Sciences de l'ingénieur [physics]

Qiong Wang, Lu Zhang, Wenbin Zou, Kidiyo Kpalma. Salient video object detection using a virtual border and guided filter. Pattern Recognition, 2020, 97, pp.106998. ⟨10.1016/j.patcog.2019.106998⟩. ⟨hal-02432462⟩