3D-GP-LMVIC: Learning-based Multi-View Image Compression with 3D Gaussian Geometric Priors

Yujun Huang, Bin Chen, Niu Lian, Xin Wang, Baoyi An, Tao Dai, Shu-Tao Xia

Keywords: Multi-View Image Compression; 3D Gaussian Splatting; Deep Learning

TL;DR: This paper introduces 3D-GP-LMVIC, a learning-based multi-view image compression method using 3D Gaussian geometric priors to improve disparity estimation and reduce redundancy.

Abstract: Multi-view image compression is vital for 3D-related applications. Existing methods often rely on 2D projection similarities between views to estimate disparity, performing well with small disparities, such as in stereo images, but struggling with more complex disparities from wide-baseline setups, common in virtual reality and autonomous driving systems. To overcome this limitation, we propose a novel approach: learning-based multi-view image compression with 3D Gaussian geometric priors (3D-GP-LMVIC). Our method leverages 3D Gaussian Splatting to derive geometric priors of the 3D scene, enabling more accurate disparity estimation between views within the compression model. Additionally, we introduce a depth map compression model to reduce redundancy in geometric information across views. A multi-view sequence ordering method is also proposed to enhance correlations between adjacent views. Experimental results demonstrate that 3D-GP-LMVIC surpasses both traditional and learning-based methods in performance, while maintaining fast encoding and decoding speed. The code is available at https://anonymous.4open.science/r/3D-GP-LMVIC-8FFA.

Primary Area: applications to computer vision, audio, language, and other modalities

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 4301