3DGS.zip: A survey on 3D Gaussian Splatting Compression Methods

• DOI: 10.48550/arxiv.2407.09510
• Publication Date: 2024-06-17
• Authors: Milena T. Bagdasa...; Paul W. Knoll; Florian Barthel; Anna Hilsmann; Peter Eisert; Wieland Morgenstern
• Keywords: FOS: Computer and information sciences; Computer Vision and Pattern Recognition (cs.CV); Computer Science - Computer Vision and Pattern Recognition

We present a work-in-progress survey on 3D Gaussian Splatting compression methods, focusing their statistical performance across various benchmarks. This aims to facilitate comparability by summarizing key statistics of different approaches in tabulated format. The datasets evaluated include TanksAndTemples, MipNeRF360, DeepBlending, and SyntheticNeRF. For each method, we report the Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Learned Perceptual Image Patch (LPIPS), resultant size megabytes (MB), as provided respective authors. is an ongoing, open project, invite contributions from research community GitHub issues or pull requests. Please visit http://w-m.github.io/3dgs-compression-survey/ for more information sortable version table.