Abstract:Detecting objects in surveillance videos is an important problem due to its wide applications in traffic control and public security. Existing methods tend to face performance degradation because of false positive or misalignment problems. We propose a novel framework, namely, Foreground Gating and Background Refining Network (FG-BR Net), for surveillance object detection (SOD). To reduce false positives in background regions, which is a critical problem in SOD, we introduce a new module that first subtracts the background of a video sequence and then generates high-quality region proposals. Unlike previous background subtraction methods that may wrongly remove the static foreground objects in a frame, a feedback connection from detection results to background subtraction process is proposed in our model to distill both static and moving objects in surveillance videos. Furthermore, we introduce another module, namely, the background refining stage, to refine the detection results with more accurate localizations. Pairwise non-local operations are adopted to cope with the misalignments between the features of original and background frames. Extensive experiments on real-world traffic surveillance benchmarks demonstrate the competitive performance of the proposed FG-BR Net. In particular, FG-BR Net ranks on the top among all the methods on hard and sunny subsets of the UA-DETRAC detection dataset, without any bells and whistles.

Automatic Foreground Seeds Discovery for Robust Video Saliency Detection

Foreground Gating and Background Refining Network for Surveillance Object Detection

Video saliency detection based on robust seeds generation and spatio-temporal propagation

Detect saliency to understand a photo

Robust Saliency Detection via Fusing Foreground and Background Priors

Video Saliency Detection Using Motion Saliency Filter

Automatic foreground extraction in video

Video Saliency Detection Using Object Proposals

Robust Salient Object Detection Via Fusing Foreground and Background Priors.

Saliency-Aware Video Object Segmentation.

Salient object detection by robust foreground and background seed selection

Saliency Detection Via Background Prior and Foreground Seeds.

Motion-Aware Rapid Video Saliency Detection

Learning Complementary Saliency Priors for Foreground Object Segmentation in Complex Scenes

Fast Video Saliency Detection Via Maximally Stable Region Motion and Object Repeatability

Saliency Detection Via Global-Object-Seed-Guided Cellular Automata

Real-time Saliency-Aware Video Abstraction.

Video-based Salient Object Detection Via Spatio-Temporal Difference and Coherence

50 FPS Object-Level Saliency Detection Via Maximally Stable Region

A Unified Video Semantics Extraction and Noise Object Suppression Network for Video Saliency Detection

Saliency-Aware Geodesic Video Object Segmentation