Sphere-GAN: a GAN-based Approach for Saliency Estimation in 360° Videos
It addresses a domain-specific problem for immersive applications by improving saliency detection in 360° videos, which is incremental as it adapts existing GAN methods to a new format.
Sphere-GAN tackles saliency estimation in 360° videos by using a GAN with spherical convolutions, outperforming state-of-the-art models on a public dataset.
The recent success of immersive applications is pushing the research community to define new approaches to process 360° images and videos and optimize their transmission. Among these, saliency estimation provides a powerful tool that can be used to identify visually relevant areas and, consequently, adapt processing algorithms. Although saliency estimation has been widely investigated for 2D content, very few algorithms have been proposed for 360° saliency estimation. Towards this goal, we introduce Sphere-GAN, a saliency detection model for 360° videos that leverages a Generative Adversarial Network with spherical convolutions. Extensive experiments were conducted using a public 360° video saliency dataset, and the results demonstrate that Sphere-GAN outperforms state-of-the-art models in accurately predicting saliency maps.