CGGS: Consistency-Augmented Geometric Gaussian Splatting for Ego-centric 3D Scene Generation

Summary · qwen2.5:32b

The paper introduces CGGS, a framework for improving ego-centric 3D scene generation by enhancing consistency and geometric accuracy through a Multi-View Latent Diffusion Model with a consistency-augmented loss and a Geometric Refiner using an entropy-based Mutual Information Depth Loss. This method outperforms previous techniques in creating coherent text-driven 3D scenes, as demonstrated through comprehensive experiments. Notably, CGGS uses optical flow and point-track correspondence to estimate depth from ego-centric 2D priors.

Suggested post angle

Learn about the new CGGS method for ego-centric 3D scene generation improving geometric structures and visual quality, enhancing real-time rendering!

Excerpt

arXiv:2607.03819v1 Announce Type: new Abstract: Challenges remain in ego-centric 3D scene generation due to limited view overlap and the dominant influence of individual perspectives on scene interpretation. These factors hinder the creation of viewpoint-consistent and semantically aligned visual content, as well as the construction of accurate geometric structures. In this paper, we propose CGGS, a text-to-3D framework aiming to enhance 3D-content-awareness and address geometric distortions in ego-centric scene generation. Firstly, the Ego-centric Generator is proposed by fine-tuning a Multi-View Latent Diffusion Model with consistency-augmented loss to generate consistent, high-fidelity 2D content aligned with textual descriptions. Then, Layout Decorator leverages optical flow and point-track correspondence to estimate depth, therefore producing dense point clouds as coarse layouts from the ego-centric 2D priors. Building on this initialization, Geometric Refiner is proposed to enhance 3D Gaussian reconstruction via an entropy-based Mutual Information Depth Loss (MID) combined with a hierarchical optimization scheme for improving visual quality and geometric structure. Comprehensive experiments demonstrate that \textcolor{softred}{CGGS} outperforms previous methods in generating coherent and accurate text-driven 3D scenes. Project page: https://cggs-26.github.io/cggs26/.