引导式测验 · 硬 等级

神经辐射场测验

回顾原始 NeRF 的神经场景表示和光线渲染，包括姿势训练图像以及密度与外观的分离。

相关引导路径神经辐射场
问题 1 的 8

Which task does the original NeRF representation support from its posed scene images?

预览每个问题8 问题

当您参加测验时，就会出现答案和解释。

  1. Which task does the original NeRF representation support from its posed scene images?
    • A guarantee of complete hidden geometry from any two photos
    • Synthesis of views not identical to the training cameras
    • Translation of every scene into a text caption
    • Mandatory face identification for every image
  2. Which quantities are predicted by the original NeRF field for queried scene coordinates?
    • RGB appearance and volume density
    • Only one complete final image returned by every point query
    • The camera’s required GPS address
    • Only the training image filename
  3. What does the conceptual five-dimensional input in the original NeRF paper combine?
    • Five separate input photographs
    • Five independent color channels
    • One date and four document labels
    • A 3D position and a viewing direction with two angular degrees of freedom
  4. How does original NeRF rendering use queried samples along a camera ray?
    • It takes only the brightest sample regardless of every other contribution
    • It ignores sample distances and every density value
    • It composites density and color with sample distances and transmittance
    • It selects an unchanged training image solely by filename
  5. What purpose does positional encoding serve in the original NeRF experiments?
    • Helping the MLP represent high-frequency coordinate-dependent detail
    • Guaranteeing a complete scene from no observations
    • Changing all spatial coordinates into camera GPS readings
    • Eliminating every optimization step
  6. 最初的 NeRF 的主要实际限制是什么？
    • 只能输出音频
    • 每个场景的训练和渲染都很慢
    • 根本不需要照片
    • 它只能在室内工作
  7. Which camera information is supplied with images in the stated original NeRF training setup?
    • Only a universal claim that all photo collections share one pose
    • Only a text prompt with no image-camera information
    • Known camera poses
    • A required list of real-world product prices
  8. In the original NeRF architecture, how is density’s dependence on input coordinates restricted relative to color?
    • Density must depend only on viewing direction and ignore position
    • Both outputs must be constant across all positions
    • Color must ignore both position and direction
    • Density uses position only; color can also use viewing direction