BEA-GS：超越 3DGS 中辐射监督的精确物体提取
BEA-GS: BEyond RAdiance Supervision in 3DGS for Precise Object Extraction
针对 3DGS 语义表示只优化外观辐射、不优化几何的问题，提出两条互补损失：一条经光栅化回传梯度修正可见高斯几何，另一条绕过光栅化直接作用于低透射率/不可见高斯，可用于物体级资产提取与场景编辑时获得干净边界。
原文依据 Our first loss propagates gradients directly through the rasterization
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论大多数提供场景三维语义表示的高斯泼溅技术并不优化场景的底层三维几何，这使得物体级编辑或资产提取变得困难。近期方法如 COBGS、Trace3D 和 ObjectGS 认识到这一局限，并提出通过修改场景几何来表示底层语义的方案。我们更进一步，提出一种新的解决方案，在物体提取中提供近乎完美的边界。为此，我们在优化中引入两个新损失，分别负责：1. 修改可见高斯的几何以遵循语义边界；2. 修改在物体被提取后才显现的不可见高斯的几何。我们的第一个损失直接通过光栅化传播梯度，从而能够无缝集成到高斯参数的优化中。第二个损失同样将梯度传播到高斯参数，但不经过光栅化。这使其即便到达某个高斯的透射率很小（部分可见或不可见）时，也能修改场景的几何。我们在 4 个数据集上使用 6 项指标与 12 种最先进方法进行了详尽比较，结果表明我们的方法产生了迄今为止整体最佳的边界分割。
英文原文摘要 1344 字符
Most Gaussian Splatting techniques that provide a 3D semantic representation of the scene don't optimize the underlying 3D geometry of the scene. This makes object-level editing or asset extraction challenging. Recent methods, like COBGS, Trace3D, and ObjectGS, acknowledge this limitation and propose approaches that modify the geometry of the scene to represent the underlying semantics. We go a step further and propose a novel solution that provides near perfect boundaries in object extraction. We do so by introducing two new losses in the optimization that take care of: 1. Modifying the geometry of visible Gaussians to respect semantic boundaries, and, 2. Modifying the geometry of non-visible Gaussians that appear once the object is extracted. Our first loss propagates gradients directly through the rasterization to allow for seamless integration within the optimization of the Gaussian parameters. Our second loss also propagates gradients to the Gaussian parameters, but does so without passing through the rasterization. This allows it to modify the geometry of the scene, even if not much transmittance arrives to a Gaussian (partial or non-visible). Exhaustive comparisons to 12 state of the art methods over 4 datasets, using six metrics, demonstrate that our approach produces overall the best boundary segmentation to date.