Scalability of B^3-Seg to larger or dynamic 3DGS scenes

Develop scalability strategies for the B^3-Seg framework to handle substantially larger or dynamic 3D Gaussian Splatting scenes by integrating approaches compatible with the analytic Expected Information Gain pipeline, enabling camera-free and training-free segmentation in broad indoor or outdoor environments.

Background

The presented method focuses on typical object-centric scenes, where analytic EIG-based active view selection and Beta–Bernoulli updates operate efficiently within a few seconds. In the discussion, the authors highlight that larger environments (e.g., wide indoor spaces or outdoor scans) may require broader viewpoint exploration strategies while remaining compatible with their EIG formulation.

They explicitly state that achieving scalability for larger or dynamic scenes, integrated into the current EIG-based pipeline, is left for future work, marking this as an unresolved extension of the framework.