3DGS-VBench: A Comprehensive Video Quality Evaluation Benchmark for 3DGS Compression

3D Gaussian Splatting (3DGS) enables real-time novel view synthesis with high visual fidelity, but its significant storage demands limit practical deployment, prompting recent methods to integrate compression modules into 3DGS. However, these 3DGS generative compression techniques introduce unique distortions that lack systematic quality assessment research. To this end, we establish 3DGS-VBench, a large-scale Video Quality Assessment (VQA) dataset and benchmark with 660 compressed 3DGS models and video sequences generated from 11 scenes across 6 representative 3DGS compression algorithms with systematically designed parameter levels. With annotations from 50 participants, we obtain MOS scores with outlier removal and validate dataset reliability. We benchmark 6 3DGS compression algorithms on storage efficiency and visual quality, and evaluate 15 quality assessment metrics. Our dataset enables specialized VQA model training for 3DGS. The dataset is available at https://github.com/YukeXing/3DGS-VBench.