# 3D Gaussian Splatting **Entity class:** Rendering method **Domain:** Computer Graphics / Neural Rendering / Scene Reconstruction **Doc Type:** Canonical Technology Node **Maturity:** Established research method **Primary source article:** [[articles/The Closed-Loop Gaussian Sensorium Engine|The Closed-Loop Gaussian Sensorium Engine]] **Collection:** [[collections/Neurotech|Neurotech]] --- ## Definition **3D Gaussian Splatting (3DGS)** is a scene-reconstruction and novel-view-synthesis method introduced by Kerbl and colleagues in 2023. It represents a scene with anisotropic three-dimensional Gaussian primitives and renders them through a visibility-aware splatting process optimized for high-quality real-time display. ## Role in the sensorium architecture [[articles/The Closed-Loop Gaussian Sensorium Engine|The Closed-Loop Gaussian Sensorium Engine]] uses 3DGS in two roles: as an external visualization method for a model's current scene belief and as a differentiable forward model that can participate in optimization. The article extends its primitive grammar—position, covariance, color, opacity-like weight, and temporal persistence—into the proposed [[wiki/Gaussian Perceptual Field|Gaussian Perceptual Field]]. ## Evidence boundary 3DGS is an established graphics method. Its use as a renderer or optimization component in a perceptual BCI is a proposed integration. Similarity between Gaussian scene primitives, receptive-field models, and phosphene maps does not establish that visual cortex literally runs the 3DGS algorithm. ## Relationships - **Rendering component of:** [[wiki/Closed-Loop Gaussian Sensorium Engine|Closed-Loop Gaussian Sensorium Engine]]. - **Machine-side representation:** [[wiki/Gaussian Perceptual Field|Gaussian Perceptual Field]]. - **Optimization target for:** [[wiki/Perceptual Attractor Seeding|Perceptual Attractor Seeding]]. - **Broader contexts:** [[wiki/Digital Twin|Digital Twin]], [[wiki/World Modeling|World Modeling]], and [[wiki/Predictive Processing|Predictive Processing]]. - **Governance consequence when coupled to neural write-paths:** [[wiki/Perceptual Sovereignty|Perceptual Sovereignty]]. ## Sources and provenance - [Kerbl, Kopanas, Leimkühler, and Drettakis, “3D Gaussian Splatting for Real-Time Radiance Field Rendering” (ACM Transactions on Graphics / SIGGRAPH 2023)](https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/) — primary paper, project page, and implementation route. - [[articles/The Closed-Loop Gaussian Sensorium Engine|The Closed-Loop Gaussian Sensorium Engine]] — source of the neurotechnology integration and cortical-analogy boundary.