SVI360: Spherical Video Interpolation

2026-07-13

AI Summary

This work addresses the performance degradation in spherical video interpolation caused by polar distortion and proposes SVI360, a novel dual-branch framework. It is the first to incorporate rotationally equivariant orthogonal views alongside the original spherical representation, leveraging the equivariance of optical flow displacement to enforce geometric consistency in polar regions. By integrating spherical video representation, a dual-branch network architecture, and a multi-view fusion strategy, the method achieves state-of-the-art performance across four public benchmarks, significantly outperforming existing approaches in both interpolation quality and optical flow accuracy.

Abstract

This paper addresses the problem of omnidirectional video interpolation, which plays an essential role in applications such as virtual reality and immersive video enhancement. Existing video interpolation methods are not well-suited for spherical videos, as they have difficulty handling severe distortions close to the poles. To address this issue, we propose SVI360, a dual-branch framework that combines the image frame and its rotated orthogonal view to deal with these distortions. The core methodological aspect of the approach is to reinforce equivariance of the flow displacements between the original and orthogonal views to improve intermediate frame prediction. Experiments show that our method outperforms state-of-the-art approaches in interpolation quality while maintaining accurate optical flow in four different public benchmarks. Code and pre-trained models are available at: https://icb-vision-ai.github.io/video360_interpolation/

Problem

• spherical video interpolation
• omnidirectional video
• pole distortion
• video frame interpolation
• virtual reality

Innovation

• spherical video interpolation
• equivariant optical flow
• dual-branch framework
• 360-degree video
• omnidirectional video