https://waymo.com/intl/es/research/block-nerf/

header
 
Ir al contenido principal
 
---------------------------------------------------------------------

  *  
    Waymo Driver
    -----------------------------------------------------------------
   
    Waymo Driver

    Como funciona nuestra tecnologia
  *  
    Waymo One
    -----------------------------------------------------------------
   
    Transportamos pasajeros

    El primer servicio de viajes autonomos del mundo
  *  
    Waymo Via
    -----------------------------------------------------------------
   
    Transportamos bienes

    Transporte autonomo y soluciones de entrega local
  *  
    Empresa
    -----------------------------------------------------------------
   
    Empresa

      +  
        Nuestra historia
      +  
        Liderazgo de Waymo
      +  
        Actualizaciones mas recientes
      +  
        Recursos de prensa
  *  
    Profesiones
    -----------------------------------------------------------------
   
    Profesiones

      +  
        Beneficios
      +  
        Valores
      +  
        Personas
      +  
        Puestos disponibles

columns
hero

Block-NeRF

Scalable Large Scene Neural View Synthesis

  *  
    Matthew Tancik*

    UC Berkeley

  *  
    Vincent Casser

    Waymo

  *  
    Xinchen Yan

    Waymo

  *  
    Sabeek Pradhan

    Waymo

  *  
    Ben Mildenhall

    Google Research

  *  
    Pratul Srinivasan

    Google Research

  *  
    Jonathan T. Barron

    Google Research

  *  
    Henrik Kretzschmar

    Waymo

* Work done as an intern at Waymo
media-grid

  * 
     
    Download high-quality video

chapter

Abstract

We present Block-NeRF, a variant of Neural Radiance Fields that can
represent large-scale environments. Specifically, we demonstrate that
when scaling NeRF to render city-scale scenes spanning multiple
blocks, it is vital to decompose the scene into individually trained
NeRFs. This decomposition decouples rendering time from scene size,
enables rendering to scale to arbitrarily large environments, and
allows per-block updates of the environment. We adopt several
architectural changes to make NeRF robust to data captured over
months under different environmental conditions. We add appearance
embeddings, learned pose refinement, and controllable exposure to
each individual NeRF, and introduce a procedure for aligning
appearance between adjacent NeRFs so that they can be seamlessly
combined. We build a grid of Block-NeRFs from 2.8 million images to
create the largest neural scene representation to date, capable of
rendering an entire neighborhood of San Francisco.

media-grid

Reconstructions of San Francisco

  * 
    Grace Cathedral  
    Download high-quality video
  * 
    Lombard Street  
    Download high-quality video
  * 
    Alamo Square Neighborhood  
    Download high-quality video
  * 
    Moscone Center (night and day)  
    Download high-quality video
  * 
    View of Bay Bridge from Embarcadero St.  
    Download high-quality video

media-grid

Appearance modulation

  * 
    Alamo Square block  
    Download high-quality video
  * 
    Downtown block  
    Download high-quality video

media-grid

Supplementary results

  * 
     
    Download high-quality video

chapter

Links

  *  
    Download PDF
  *  
    arXiv

Copy BibTeX
footer
 

  *  
    Preguntas frecuentes
  *  
    Blog
  *  
    Politica de privacidad
  *  
    Conjunto de datos abierto
  *  
    Condiciones
  *  
    Legal
  *  
    Tolerancia cero
  *  
    Servicios de primera respuesta
  *  
    Informe de seguridad
  *  
    Hablemos sobre la conduccion autonoma
  *  
    Contactarse
  *  
    Recursos ante el COVID-19

Registrate para recibir actualizaciones acerca de las ultimas
noticias sobre Waymo y nuestra tecnologia.
 
Registro
[Espanol         ]
---------------------------------------------------------------------

  *  
  *  
  *  
  *  
  *  

(c) 2019-2022 Waymo LLC