360-Degree Video Super Resolution and Quality Enhancement Challenge: Methods and Results

TL;DR

This paper presents a challenge focused on developing machine learning methods for super-resolution and quality enhancement of 360-degree videos, aiming to improve real-time streaming quality under bandwidth and latency constraints.

Contribution

It introduces a new challenge framework for 360-degree video super-resolution, detailing competition tracks, evaluation criteria, and analyzing top solutions to advance immersive video quality.

Findings

Top models achieved significant bitrate gains with enhanced quality.

Proposed methods improved super-resolution efficiency for 360-degree videos.

Framework facilitates benchmarking of real-time quality enhancement solutions.

Abstract

Omnidirectional (360-degree) video is rapidly gaining popularity due to advancements in immersive technologies like virtual reality (VR) and extended reality (XR). However, real-time streaming of such videos, especially in live mobile scenarios like unmanned aerial vehicles (UAVs), is challenged by limited bandwidth and strict latency constraints. Traditional methods, such as compression and adaptive resolution, help but often compromise video quality and introduce artifacts that degrade the viewer experience. Additionally, the unique spherical geometry of 360-degree video presents challenges not encountered in traditional 2D video. To address these issues, we initiated the 360-degree Video Super Resolution and Quality Enhancement Challenge. This competition encourages participants to develop efficient machine learning solutions to enhance the quality of low-bitrate compressed 360-degree videos, with two tracks focusing on 2x and 4x super-resolution (SR). In this paper, we outline the challenge framework, detailing the two competition tracks and highlighting the SR solutions proposed by the top-performing models. We assess these models within a unified framework, considering quality enhancement, bitrate gain, and computational efficiency. This challenge aims to drive innovation in real-time 360-degree video streaming, improving the quality and accessibility of immersive visual experiences.