Summary · Chapter 17

Feature fields

Distilled feature field: a 3D function trained so its rendering from each camera matches a 2D teacher's feature map.

WorkKey factVersus this project
Distilled Feature Fields, LERFCLIP, DINO or LSeg features distilled into a NeRF; text, patch or click queriesImplicit, 3D, per scene
Feature 3DGSArbitrary-dimension features on 3D Gaussians (SAM, LSeg)3D, per scene, optimised features
LangSplatCLIP features via a scene-wise autoencoder; 199x faster than LERF at 1440x1080Compresses features; yours keeps 384-d
Upsample AnythingOne Gaussian kernel per coarse pixel; 0.419 s per 224 px imageFilter on the grid; features not on Gaussians
This project2D Gaussians own features; closed-form solve; amortised placementSingle image, any count

Defence answer. “Feature 3DGS puts optimised features on 3D Gaussians per scene. I put closed-form features on 2D Gaussians placed by a trained network for one image, at any count. Upsamplers produce a dense grid; mine is a compact set read at any location. A matched-protocol comparison with upsamplers is future work.”

Caveats: semantics are inherited from the teacher, which is still needed at inference; no fair upsampler comparison yet.