A novel superpixel-based saliency detection model for 360-degree images

作者：

Highlights：

• We propose a saliency detection model for 360-degree images by figure-ground law of Gestalt theory with feature contrast and boundary connectivity.

• We design a novel background measure called 360-degree boundary connectivity, which is reliable to compute background prior in 360-degree images.

• Different from some existing 360-degree saliency detection models, we directly use original images instead of other format by projection to avoid introducing noise or missing information.

摘要

•We propose a saliency detection model for 360-degree images by figure-ground law of Gestalt theory with feature contrast and boundary connectivity.•We design a novel background measure called 360-degree boundary connectivity, which is reliable to compute background prior in 360-degree images.•Different from some existing 360-degree saliency detection models, we directly use original images instead of other format by projection to avoid introducing noise or missing information.

论文关键词：Visual attention,360-degree image,Saliency detection,Figure-ground law,Boundary connectivity,Gestalt theory

论文评审过程：Received 11 September 2017, Revised 25 July 2018, Accepted 25 July 2018, Available online 1 August 2018, Version of Record 30 October 2018.

论文官网地址：https://doi.org/10.1016/j.image.2018.07.009