Saliency Prediction for Omnidirectional Images Considering Optimization on Sphere Domain

TL;DR

This paper introduces a novel method for predicting saliency maps in omnidirectional images by optimizing on the sphere domain, effectively adapting 2D saliency predictors to ERP images with pre- and post-processing techniques.

Contribution

It presents a new approach that applies sphere domain optimization to extend 2D saliency prediction methods to omnidirectional ERP images.

Findings

Achieves competitive saliency prediction results on a public omnidirectional image dataset.

Utilizes sphere domain smoothing to improve prediction accuracy.

Effectively manages ERP image distortions through pre- and post-processing.

Abstract

There are several formats to describe the omnidirectional images. Among them, equirectangular projection (ERP), represented as 2D image, is the most widely used format. There exist many outstanding methods capable of well predicting the saliency maps for the conventional 2D images. But these works cannot be directly extended to predict the saliency map of the ERP image, since the content on ERP is not for direct display. Instead, the viewport image on demand is generated after converting the ERP image to the sphere domain, followed by rectilinear projection. In this paper, we propose a model to predict the saliency maps of the ERP images using existing saliency predictors for the 2D image. Some pre-processing and post-processing are used to manage the problem mentioned above. In particular, a smoothing based optimization is realized on the sphere domain. A public dataset of omnidirectional images is used to perform all the experiments and competitive results are achieved.