用于复杂场景背景减除的拮抗方向能量时空显著性

本文基于生物视觉启发构建了一种用于复杂场景下的无监督时空显著性检测方法。为了紧致描述场景局部动态纹理运动,首先提出分布式拮抗方向能量导出的动态纹理描述子获取局部运动信息。然后,本文探索了利用该表达结合扩展的自相似显著性框架得到时空显著性检测器。在多个极端动态背景和摄像头运动的图像序列中,利用本文提出的算法进行了背景减除实验。广泛的定性与定量实验结果表明本文方法快速且高效。在检测精度与计算代价平衡性方面,本文算法与最先进算法具有一定竞争力。

Spatiotemporal saliency based on opponent oriented energy for background subtraction in complex scenes

A computational unsupervised saliency model utilizing bio-inspired features for spatiotemporal saliency is presented in this paper.We first propose distributed opponent oriented energy for capturing local dynamic texture ensemble motion.Then,we integrate the derived characterization and a revised self-resemblance saliency framework.High effectiveness and efficiency of the proposed method is extensively demonstrated both qualitatively and quantitatively,for background subtraction in the cases of extremely dynamic scenes and camera jitter.In terms of the trade-off between accuracy and computation cost,our method achieved competitive results in contrast to the state-of-art algorithm.