Cogs dataset reveals varied usefulness of 3D Gaussian attributes
COGS: A Diagnostic Benchmark for Attribute-Aware Object-Level 3D Gaussian Splatting Perception
Computer Vision and Pattern Recognition
Summary
3D Gaussian splatting is a way to represent 3D objects that includes multiple attributes, but it’s unclear how these attributes affect object recognition. The paper’s authors created a large, detailed dataset called COGS, with many objects having diverse appearances, to test how these attributes impact perception. They found that not all attributes help equally; sometimes using fewer can work better. They also designed a model named HALO to keep attribute information organized, improving understanding of how these attributes are used.
What this means in practice
- •For 3d vision engineers: Evaluate and optimize how 3D Gaussian attributes contribute to object recognition models using the COGS dataset and benchmark.
- •For augmented reality developers: Improve 3D object representation in AR by using attribute-specific insights from COGS to enhance perception accuracy.
Authors
Zheng Xiao, Yifan Liu, Weijian Deng, Jiashu Li, Ziqin Liu, Zhenjun Han, Jianbin Jiao
Abstract
3D Gaussian Splatting (3DGS) is increasingly explored as an input representation for 3D perception. Yet it remains unclear how current 3D architectures exploit heterogeneous Gaussian attributes. This ambiguity makes accuracy hard to interpret and limits principled architecture design for 3DGS perception. However, existing object-level 3DGS datasets lack the appearance variation or artifact control needed for systematic attribute-level analysis. We introduce COGS, a large-scale 3DGS dataset with 15,675 artifact-filtered and appearance-rich objects across 40 categories. Built on COGS, we establish COGS-Bench for object classification and attribute-usage diagnosis, complemented by external validation. Our benchmark shows that Gaussian attributes are not uniformly beneficial: selected subsets can outperform the full attribute set, and the effects of appearance, opacity, and covariance-related attribute combinations vary substantially across architectures. We further study this finding with HALO, a reference model that preserves attribute-specific evidence before hierarchical aggregation. Experiments on COGS-Bench and external 3DGS datasets demonstrate that COGS supports systematic diagnosis of 3DGS attribute usage, while HALO highlights the value of role-preserving attribute organization. Our dataset is available at https://huggingface.co/datasets/XiaoZhengCV/COGS.