GS-QA: Comprehensive Quality Assessment Benchmark for Gaussian Splatting View Synthesis

TL;DR

This paper introduces a comprehensive benchmark for assessing the quality of Gaussian Splatting view synthesis, including subjective and objective evaluations across diverse scenes and camera trajectories.

Contribution

It presents the first extensive subjective quality assessment and benchmarks 18 objective metrics for Gaussian Splatting view synthesis, filling a gap in quality evaluation methods.

Findings

Subjective scores reveal strengths and limitations of GS methods.

Objective metrics vary in alignment with human perception.

The benchmark database supports future research in GS quality assessment.

Abstract

Gaussian Splatting (GS) offers a promising alternative to Neural Radiance Fields (NeRF) for real-time 3D scene rendering. Using a set of 3D Gaussians to represent complex geometry and appearance, GS achieves faster rendering times and reduced memory consumption compared to the neural network approach used in NeRF. However, quality assessment of GS-generated static content is not yet explored in-depth. This paper describes a subjective quality assessment study that aims to evaluate synthesized videos obtained with several static GS state-of-the-art methods. The methods were applied to diverse visual scenes, covering both 360-degree and forward-facing (FF) camera trajectories. Moreover, the performance of 18 objective quality metrics was analyzed using the scores resulting from the subjective study, providing insights into their strengths, limitations, and alignment with human perception. All videos and scores are made available providing a comprehensive database that can be used as benchmark on GS view synthesis and objective quality metrics.