Search papers, labs, and topics across Lattice.
This paper addresses the problem of artifacts in sparse-view 3D Gaussian Splatting (3DGS) reconstruction, which arise from insufficiently constrained Gaussians. They propose a Dual-domain Observation and Calibration (DOC-GS) framework that uses a Continuous Depth-Guided Dropout (CDGD) in the optimization domain to represent Gaussian reliability and a Dark Channel Prior (DCP) in the observation domain to identify and remove unreliable Gaussians. Experiments demonstrate that DOC-GS effectively reduces artifacts and improves reconstruction quality in sparse-view scenarios.
Unreliable Gaussians are the root cause of artifacts in sparse-view 3D Gaussian Splatting, and modeling their reliability in both the optimization and observation domains dramatically improves reconstruction quality.
Sparse-view reconstruction with 3D Gaussian Splatting (3DGS) is fundamentally ill-posed due to insufficient geometric supervision, often leading to severe overfitting and the emergence of structural distortions and translucent haze-like artifacts. While existing approaches attempt to alleviate this issue via dropout-based regularization, they are largely heuristic and lack a unified understanding of artifact formation. In this paper, we revisit sparse-view 3DGS reconstruction from a new perspective and identify the core challenge as the unobservability of Gaussian primitive reliability. Unreliable Gaussians are insufficiently constrained during optimization and accumulate as haze-like degradations in rendered images. Motivated by this observation, we propose a unified Dual-domain Observation and Calibration (DOC-GS) framework that models and corrects Gaussian reliability through the synergy of optimization-domain inductive bias and observation-domain evidence. Specifically, in the optimization domain, we characterize Gaussian reliability by the degree to which each primitive is constrained during training, and instantiate this signal via a Continuous Depth-Guided Dropout (CDGD) strategy, where the dropout probability serves as an explicit proxy for primitive reliability. This imposes a smooth depth-aware inductive bias to suppress weakly constrained Gaussians and improve optimization stability. In the observation domain, we establish a connection between floater artifacts and atmospheric scattering, and leverage the Dark Channel Prior (DCP) as a structural consistency cue to identify and accumulate anomalous regions. Based on cross-view aggregated evidence, we further design a reliability-driven geometric pruning strategy to remove low-confidence Gaussians.
