Feed-Forward Gaussian Splatting from Asymmetric Dual-lens Stereo

Abstract

Modern devices commonly deploy multiple cameras, such as dual-lens camera systems on smartphones. Because the field-of-views differ between the cameras, image resolutions in overlapping regions can vary. Unfortunately, feed-forward 3D Gaussian Splatting (3DGS) methods are designed only for images with the same resolutions, which limits their performance with dual-lens images. We observe that existing works struggle to preserve high-frequency details and to account for 3D geometry from images with different resolutions and narrow baselines. In this paper, we propose a novel feed-forward 3DGS method that handles dual-lens inputs. To do this, we first consider a dual-lens camera system as a cross-scale input. We then adopt a coarse-to-fine feature encoding strategy to enhance low-resolution image features by leveraging geometric information from a depth foundation model. We improve low-resolution images to provide high-frequency details using a cross-scale enhancement module. Finally, we predict Gaussian parameters through a decoding unit to yield 3D Gaussian primitives. Our model achieves state-of-the-art performance on various datasets.