# Video Salient Object Detection via Fully Convolutional Networks

## TL;DR

This paper introduces a deep learning framework for video saliency detection that efficiently captures spatial and temporal cues, utilizing synthetic data augmentation to overcome limited annotated video datasets, and achieves state-of-the-art accuracy with real-time speed.

## Contribution

The paper presents a novel deep video saliency network with separate modules for spatial and temporal information, and a data augmentation technique to enhance training with limited video data.

## Key findings

- Achieved state-of-the-art performance on DAVIS and FBMS datasets.
- Produced accurate spatiotemporal saliency maps at 2 fps.
- Effectively used synthetic data to improve model robustness.

## Abstract

This paper proposes a deep learning model to efficiently detect salient regions in videos. It addresses two important issues: (1) deep video saliency model training with the absence of sufficiently large and pixel-wise annotated video data, and (2) fast video saliency training and detection. The proposed deep video saliency network consists of two modules, for capturing the spatial and temporal saliency information, respectively. The dynamic saliency model, explicitly incorporating saliency estimates from the static saliency model, directly produces spatiotemporal saliency inference without time-consuming optical flow computation. We further propose a novel data augmentation technique that simulates video training data from existing annotated image datasets, which enables our network to learn diverse saliency information and prevents overfitting with the limited number of training videos. Leveraging our synthetic video data (150K video sequences) and real videos, our deep video saliency model successfully learns both spatial and temporal saliency cues, thus producing accurate spatiotemporal saliency estimate. We advance the state-of-the-art on the DAVIS dataset (MAE of .06) and the FBMS dataset (MAE of .07), and do so with much improved speed (2fps with all steps).