Cuestionario vinculado a una guía · duro Nivel

Prueba de salpicaduras gaussianas

Estudie las primitivas y el renderizador explícitos del 3D Gaussian Splatting original, junto con la optimización, la inicialización, el color direccional y la covarianza.

Rutas de guía relacionadasSalpicadura gaussiana
Pregunta 1 de 8

Which representation defines a scene in the original 3D Gaussian Splatting method?

Vista previa de cada pregunta8 preguntas

Las respuestas y explicaciones aparecen a medida que realiza el cuestionario.

  1. Which representation defines a scene in the original 3D Gaussian Splatting method?
    • A required single hidden MLP with no explicit primitives
    • An explicit set of anisotropic 3D Gaussian primitives
    • Only one flat reference photograph
    • A text description with no spatial parameters
  2. Which rendering procedure is central to the original Gaussian Splatting method?
    • Projection and visibility-aware tile-based alpha blending
    • Only repeated MLP queries along every original NeRF ray
    • A rule that all primitives bypass training
    • Only a black-and-white image average
  3. Which property set is modeled for each primitive in the original method?
    • Only an output filename
    • Only a scalar volume-density field for the entire scene
    • Only one mandatory class label for all photographs
    • Position, covariance, opacity, and appearance
  4. ¿Qué se utiliza para permitir que el color de una mancha cambie con el ángulo de visión?
    • Compresión JPEG
    • Solo un valor RGB fijo
    • Armónicos esféricos
    • Ruido aleatorio
  5. Which initialization supplies sparse points for the stated original method?
    • Structure from motion
    • A guaranteed complete manually built mesh
    • Only a text-to-image prompt
    • A required globally identical random point set
  6. When was the cited original 3D Gaussian Splatting research presented?
    • 2015
    • 2023
    • 2020
    • 2026
  7. In the paper’s adaptive density control, which operation can add coverage from a small Gaussian in an under-reconstructed region?
    • Deleting every observed image
    • Forcing every primitive to be one isotropic scene-wide blob
    • Cloning it into an additional primitive
    • Replacing its appearance with a mandatory text label
  8. Why parameterize the original Gaussian covariance through scale and valid rotation factors?
    • To guarantee that no scene artifact can occur
    • To turn every Gaussian into a text token
    • To remove all direction-dependent appearance parameters
    • To preserve a valid positive-semidefinite covariance form