Semi-Supervised Video Salient Object Detection Using Pseudo-Labels

Deep learning-based video salient object detection has recently achieved great success with its performance significantly outperforming any other unsupervised methods. However, existing data-driven approaches heavily rely on a large quantity of pixel-wise annotated video frames to deliver such promising results. In this paper, we address the semi-supervised video salient object detection task using pseudo-labels. Specifically, we present an effective video saliency detector that consists of a spatial refinement network and a spatiotemporal module. Based on the same refinement network and motion information in terms of optical flow, we further propose a novel method for generating pixel-level pseudo-labels from sparsely annotated frames. By utilizing the generated pseudo-labels together with a part of manual annotations, our video saliency detector learns spatial and temporal cues for both contrast inference and coherence enhancement, thus producing accurate saliency maps. Experimental results demonstrate that our proposed semi-supervised method even greatly outperforms all the state-of-the-art fully supervised methods across three public benchmarks of VOS, DAVIS, and FBMS.

PDF Abstract ICCV 2019 PDF ICCV 2019 Abstract

Results from the Paper

 Ranked #1 on Video Salient Object Detection on VOS-T (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Video Salient Object Detection DAVIS-2016 RCRNet+NER S-Measure 0.884 # 4
MAX F-MEASURE 0.859 # 4
AVERAGE MAE 0.028 # 7
Video Salient Object Detection FBMS-59 RCRNet+NER S-Measure 0.870 # 3
AVERAGE MAE 0.054 # 4
MAX F-MEASURE 0.861 # 3
Video Salient Object Detection VOS-T RCRNet+NER S-Measure 0.872 # 1
max E-measure 0.856 # 1
Average MAE 0.049 # 1


No methods listed for this paper. Add relevant methods here